CN101452472A - Information processing device, method, and program - Google Patents

Abstract

An information processing device, a method and program are disclosed. The information processing device includes an item evaluation acquiring section configured to acquire evaluation values given to individual items by individual users, a user statistics calculating section configured to calculate user statistics indicating an evaluation tendency of a noted user, by using at least one of the number of items evaluated by the noted user, evaluation values given by the noted user to individual items, the numbers of evaluations given by individual users to items evaluated by the noted user, and evaluation values given by individual users to items evaluated by the noted user, and a presentation control section configured to control presentation of information related to an item to the noted user, on the basis of the user statistics.

Description

Messaging device and method and program

The cross reference of related application

The present invention comprise with respectively on Dec 3rd, 2007 and on July 2nd, 2008 the Japanese patent application JP 2007-312722 theme relevant with JP 2008-173489 to the application of Jap.P. office, by reference its whole contents is incorporated into here.

Technical field

The present invention relates to a kind of messaging device and method and program.More particularly, the present invention relates to a kind of messaging device and method and the program that can more effectively use user's evaluation of the project of giving.

Background technology

In correlation technique, various inventions have been proposed for so-called content personalization, wherein take out such as TV programme, the multistage music, and such various projects of works and recommend (referring to the open 2004-194107 of for example Japanese Unexamined Patent Application according to user's preference, perhaps P.Resnick, N.Iacovou, M.Suchak, P.Bergstrom, and " GroupLens:Open Architecture for Collaborative FilterIng of Netnews " (Conference on Computer Supported Cooperative Work of J.RIedl, the 175-186 page or leaf, 1994)).For content personalization, be extensive use of such as the cooperation of estimating based on the user and filter (CF) and based on the such method of content-based filtration (CBF) of the information content.

Summary of the invention

For estimating for the situation of the recommended project filtering or the like according to the user by cooperation in the correlation technique, in order to allow to recommend more suitable project, expectation can more effectively use the user of the project of giving to estimate.

Therefore desirable is more effectively to use the user of the project of giving to estimate.

Messaging device according to the embodiment of the invention comprises: the project evaluation deriving means is used to obtain the evaluation of estimate that each user gives each project; User's statistical computation device is used for by use paying close attention to number of items that the user estimated, paying close attention to evaluation of estimate, each user that the user gives each project pay close attention at least one of evaluation of estimate of the project that the user estimated of the evaluation number of the project that the user estimated and each user of paying close attention and calculate user's statistics of paying close attention to evaluation of user trend; And present control device, be used for adding up to control the information relevant with project presented to paying close attention to the user according to the user.

This messaging device may further include the project Extension arrangement that is used for making by the use preordering method project cluster, and user's statistical computation device can calculate user's statistics according to the cluster specific distribution of paying close attention to the number of items that the user estimated.

User statistics can comprise that group represent index, and on behalf of exponential representation, this group pay close attention to index of similarity between the cluster specific distribution of evaluation number of the affiliated whole group of the cluster specific distribution of the number of items that the user estimated and concern user.

User's statistics may further include the modish index (trendiness index) of representing the sequential mean value of index based on group.

User statistics can comprise index of conformity, and this index of conformity is paid close attention to the sequential stability index of the cluster specific distribution of the number of items that the user estimates.

User's statistics can comprise the prejudice index, and this prejudice exponential representation is in the prejudice degree of the cluster specific distribution of paying close attention to the number of items that the user estimated.

Present control device and can add up the information that represented concern user's feature is complementary presenting to control so that select and present with the user.

This messaging device may further include project statistical computation device, this project statistical computation device be used for the evaluation of estimate that gives according to each user and estimate number at least one calculate the project statistics that is used to represent give the evaluation trend of each project.

User's statistical computation device can according to project add up in the middle of the represented item characteristic, pay close attention to the feature that bulk items had that the user estimated and calculate user's statistics of paying close attention to the user.

The project statistics can comprise instantaneous index, index passes from mouth to mouth, and in the designation number at least one, described instantaneous index is based on but underspeeding of the evaluation number of each each project estimated the relative value of the average velocity that number reduces with respect to become the time spent from each project, time segment length and the increase degree of estimating number that the described exponential representation of passing from mouth to mouth increases the evaluation number of each each project, described designation number is represented the sequential stability index to the evaluation number of each each project, and user's statistics can comprise faddict's index, connoisseur's index, and in the conservative index at least one, described faddict's index is based on being estimated within the scheduled time slot that becomes in project after available, and each has equaling or is higher than the project and the ratio of paying close attention to the project that the user estimated of the instantaneous index of predetermined threshold, described connoisseur's index is based on being estimated within the scheduled time slot that becomes in project after available, and each has equaling or is higher than the project and the ratio of paying close attention to the project that the user estimated of the index that passes from mouth to mouth of predetermined threshold, and described conservative index is each ratio of project and the project that the user estimated of concern that has equaling or be higher than the designation number of predetermined threshold based on it.

Project statistics can comprise based on each user within scheduled time slot to the fixing fan's index of the project of the average ratings number of each each project, and user statistics can comprise fixedly fan's index of user, this user fixedly fan's index based on it the fixedly project of fan's index and ratio of the concern project that the user estimated of each project that has equaling or be higher than predetermined threshold.

Project statistics can comprise based on the main index to the evaluation number of each each project, and as the evaluation mean value of the mean value of the evaluation of estimate of each each project, and user's statistics can comprise faddict's index, main orientation index, general index number, and reputation orientation index, described faddict's index is based on the mean value of the main index of each each project that the user estimated of concern, described main orientation index is based on the correlativity of paying close attention between the main index that the user gives the evaluation of estimate of each each project and this project, described general index number is based on the correlativity between the evaluation mean value of paying close attention to evaluation of estimate that the user gives each each project and this project, and described reputation orientation index is based on the mean value of the evaluation mean value of concern each each project that the user estimated.

Present control device can make project statistics represented and add up the relevant highlighted demonstration of item characteristic of represented concern user's feature with the user and present.

Messaging device may further include the extraction element that is used to extract following project, described project have project statistics represented and add up represented concern user's the relevant feature of feature with the user, and present control device and can present this and control so that the project of being extracted is presented to the concern user.

Messaging device may further include: user's index of similarity calculation element is used for adding up the user's index of similarity that calculates the index of similarity between the expression user according to the user; The similar users extraction element is used to extract the similar users similar to paying close attention to the user; And extraction element, be used to extract project that similar users gives its high evaluation value with as will recommending the project of paying close attention to the user, and present control device and can present this and control so that the project of being extracted is presented as the project that will recommend the concern user.

Messaging device may further include: user's index of similarity calculation element is used for adding up the user's index of similarity that calculates the index of similarity between the expression user according to the user; The prediction and evaluation value calculation apparatus, be used for by use other users pay close attention the evaluation of estimate of project, by distributing big weighting to the high evaluation of estimate that the user gave of itself and the value of user's index of similarity of paying close attention to the user and paying close attention to the pay close attention predicted value of evaluation of estimate of project of user by distributing little weighting to the low evaluation of estimate that the user gave of itself and the value of user's index of similarity of paying close attention to the user, calculating; And extraction element, be used to extract the high project of the evaluation of estimate of prediction with as will recommending the project of paying close attention to the user, and present control device and can present this and control so that the project of being extracted is presented as the project that will recommend the concern user.

Information processing method may further comprise the steps according to an embodiment of the invention: obtain the evaluation of estimate that each user gives each project; Calculate user's statistics of paying close attention to evaluation of user trend by at least one that use to pay close attention to number of items that the user estimated, pay close attention to that evaluation of estimate, each user that the user gives each project pay close attention that the evaluation number of the project that the user estimated and each user pay close attention in the evaluation of estimate of the project that the user estimated; And add up to control the information relevant with project presented to according to the user and pay close attention to the user.

Program can make computing machine carry out the processing that comprises the steps according to an embodiment of the invention: obtain the evaluation of estimate that each user gives each project; Calculate user's statistics of paying close attention to evaluation of user trend by at least one that use to pay close attention to number of items that the user estimated, pay close attention to that evaluation of estimate, each user that the user gives each project pay close attention that the evaluation number of the project that the user estimated and each user pay close attention in the evaluation of estimate of the project that the user estimated; And add up to control the information relevant with project presented to according to the user and pay close attention to the user.

According to embodiments of the invention, obtain the evaluation of estimate that each user gives each project, calculate user's statistics of paying close attention to evaluation of user trend by at least one that use to pay close attention to number of items that the user estimated, pay close attention to that evaluation of estimate, each user that the user gives each project pay close attention that the evaluation number of the project that the user estimated and each user pay close attention in the evaluation of estimate of the project that the user estimated, and add up to control the information relevant with project presented to according to the user and pay close attention to the user.

According to embodiments of the invention, can more effectively use the user to give the evaluation of project.Especially, according to embodiments of the invention, the information relevant with project suitably can be presented to the user.

Description of drawings

Fig. 1 is the block scheme that has provided according to the information handling system of the embodiment of the invention;

Fig. 2 is used for the process flow diagram that processing describes is obtained in project evaluation;

Fig. 3 is the synoptic diagram that has provided the example of project evaluation history;

Fig. 4 is used for the process flow diagram that computing describes to item characteristic;

Fig. 5 is the synoptic diagram that has provided the example of project statistics;

Fig. 6 is the synoptic diagram that has provided the rank example in the project statistics;

Fig. 7 is the synoptic diagram that has provided the example of item types index;

Fig. 8 is used for similar terms is extracted the process flow diagram that processing describes;

Fig. 9 is the synoptic diagram that has provided the example of project index of similarity;

Figure 10 is used for the process flow diagram that computing describes to user characteristics;

Figure 11 is the synoptic diagram that has provided the example of user's statistics;

Figure 12 is the synoptic diagram that has provided the example of relative faddict's index;

Figure 13 is used for similar users is extracted the process flow diagram that processing describes;

Figure 14 is the synoptic diagram that has provided the example of distance and user's index of similarity between the user;

Figure 15 is used for the process flow diagram that describes is handled in project recommendation;

Figure 16 is used for process flow diagram that second embodiment that project recommendation is handled is described;

Figure 17 has gathered to be used to calculate the form of formula of each index of type of identifying project;

Figure 18 is the evaluation mean value that has gathered project, the form of estimating the relation between variance and evaluation number and each item types;

Figure 19 is the block scheme that has provided according to the information handling system of second embodiment of the invention;

Figure 20 is used for the process flow diagram that computing describes to user characteristics (reputation orientation index);

Figure 21 is used for the process flow diagram that computing describes to user characteristics (most orientation index);

Figure 22 has provided the synoptic diagram that the user is categorized into the example of user's cluster;

Figure 23 is used for the process flow diagram that computing describes to user characteristics (prejudice index);

Figure 24 has provided the synoptic diagram that classification of the items is become the example of project cluster;

Figure 25 has provided the synoptic diagram of example of the number of items that the user estimated being made the result of form according to the project cluster;

Figure 26 has provided the synoptic diagram of another example of the number of items that the user estimated being made the result of form according to the project cluster;

Figure 27 is used for the process flow diagram that computing describes to user characteristics (group represents index);

Figure 28 has provided the synoptic diagram of all evaluation of user total numbers being made the example as a result of form according to the project cluster;

Figure 29 is used for the process flow diagram that computing describes to user characteristics (index of conformity/current old index of modish index/oneself);

Figure 30 be provided according to the project cluster, the user estimates the synoptic diagram of the example that time that number distributes changes;

Figure 31 be provided broken according to the project cluster, the user estimates the synoptic diagram of another example that time that number distributes changes;

Figure 32 be provided broken according to the project cluster, the user estimates the synoptic diagram of the another example that time that number distributes changes;

Figure 33 be provided broken according to the project cluster, the user estimates the synoptic diagram of the example that time that total number distributes changes;

Figure 34 is used for the process flow diagram that computing describes to item characteristic (the instantaneous index/index/designation number that passes from mouth to mouth/fixedly fan's index);

Figure 35 has provided in each relative period the evaluation number of project to be made the result's of form the synoptic diagram of example;

Figure 36 has provided the synoptic diagram that calculates the result of the evaluation number for the previous period at the tabulating result of the project among Figure 35;

Figure 37 is the synoptic diagram that has provided the example that the time to the evaluation number of instantaneous type project changes;

Figure 38 is the synoptic diagram that has provided the example that the time to the evaluation number of the type project of passing from mouth to mouth changes;

Figure 39 is the synoptic diagram that has provided the example that the time to the evaluation number of standard form project changes;

Figure 40 has provided the synoptic diagram of each each user to the example of the time transformation of the evaluation number of project;

Figure 41 has provided the synoptic diagram of each each user to another example of the time transformation of the evaluation number of project;

Figure 42 is used for the process flow diagram that computing describes to user characteristics (faddict B index/connoisseur's index/conservative index/fixedly fan's index);

Figure 43 is the form that has gathered item characteristic;

Figure 44 is the form that has gathered user characteristics;

Figure 45 is used for process flow diagram that the message block personalisation process is described;

Figure 46 has provided in the music distribution business synoptic diagram of example that is shown to user's screen by the message block personalisation process;

Figure 47 has provided in the music distribution business synoptic diagram of another example that is shown to user's screen by the message block personalisation process;

Figure 48 is used for process flow diagram that filtration treatment is described;

Figure 49 is used for the synoptic diagram that the concrete example to filtration treatment describes;

Figure 50 is used for process flow diagram that the highlighted display process of item characteristic is described;

Figure 51 is the synoptic diagram of example that is shown to user's screen in the music distribution business by the highlighted display process of item characteristic;

Figure 52 is used for process flow diagram that popular (hit) prediction processing is described;

Figure 53 is the synoptic diagram of example that has provided the owning rate of user characteristics;

Figure 54 is the synoptic diagram of another example that has provided the owning rate of user characteristics; And

Figure 55 is the synoptic diagram that has provided the example of computer configuration.

Embodiment

Hereinafter, with reference to the accompanying drawings embodiments of the invention are described.

Fig. 1 has provided the block scheme of information handling system according to an embodiment of the invention.Information handling system 1 among Fig. 1 be used for to the user provide project, the information relevant with project, with the system of subscriber-related information of information handling system 1 or the like.Employed here term " project " is meant such as the so all kinds of contents of TV programme, moving image, rest image, document, multistage music, software and information, various works or the like.Information handling system 1 comprises user interface part 11 and information processing part 12.

When the user is input to information processing part 12 with information or order, perhaps when the project that information processing part 12 is provided or information are presented to the user, use user interface part 11.User interface part 11 comprises by keyboard, mouse or the like importation 21 of forming of configuration and by being included in the display part 22 that display within CE (consumer electronics) equipment or the like configuration forms.

Information processing part 12 comprises that project evaluation obtains part 31, historical retaining part 32, project statistical computation part 33, item types determining section 34, project index of similarity calculating section 35, similar terms extracts part 35, user's statistical computation part 37, user's index of similarity calculating section 38, similar users is extracted part 39, prediction and evaluation value calculating section 40, the recommended project is extracted part 41, information presents part 42, project information retaining part 43, and user profile retaining part 44.

Project evaluation is obtained part 31 and is obtained following information and the information of being obtained is recorded the project evaluation that remains in the historical retaining part 32 in history, the evaluation to each project that each user of wherein said information representation is imported by importation 21.

As described with reference to figure 4 or the like subsequently, the history of project information that project statistical computation part 33 is kept according to historical retaining part 32 is calculated the item statistics that is used to represent to the evaluation trend of each project.Project statistical computation part 33 will be used to as required to represent that the information of the project statistics calculated offers item types determining section 34, project index of similarity calculating section 35 and user's statistical computation part 37.

As described with reference to figure 4 or the like subsequently, item types determining section 34 is determined the item type of the feature of each each project of expression according to the evaluation trend that gives each project.This item types determining section 34 will represent that the information of the item types of each project offers information and presents part 42.

As described with reference to figure 8 or the like subsequently, project index of similarity calculating section 35 calculates the project index of similarity of the index of similarity that is used for the central evaluation trend of expression project.Project index of similarity calculating section 35 will be used to represent that the information of the project index of similarity that calculated offers similar terms and extracts part 36.

As described with reference to figure 8 or the like subsequently, with regard to each project, similar terms extracts part 36 and extracts the similar terms similar to project according to the project index of similarity.Similar terms extracts information that part 36 will be used to represent the similar terms of each project and offers information and present part 42.

As described with reference to Figure 10 or the like subsequently, user's statistical computation part 37 according to project evaluation historical and project add up and calculate following user and add up, described user's statistical representation is based on each user's of the evaluation trend that gives each project feature.This user's statistical computation part 37 will be used to as required to represent that the information of user's statistics of being calculated offers user's index of similarity calculating section 38 and information presents part 42.

As described with reference to Figure 13 or the like subsequently, user's index of similarity calculating section 38 is added up according to the user and is calculated the user's index of similarity that is used to represent the index of similarity in the middle of the user.User's index of similarity calculating section 38 will be used to as required to represent that the information of user's index of similarity of being calculated offers similar users and extracts part 39 and prediction and evaluation value calculating section 40.

As described with reference to Figure 13 or the like subsequently, similar users is extracted part 39 and is extracted the similar users similar to each user according to user's index of similarity.This similar users is extracted information that part 39 will be used to represent each user's similar users as required and is offered that the recommended project is extracted part 41 and information presents part 42.

As described with reference to Figure 15 or the like subsequently, predictor calculation part 40 is calculated following prediction and evaluation value, and this prediction and evaluation value is the predicted value to the evaluation of estimate of the unvalued project of user.This prediction and evaluation value calculating section 40 will be used to represent that the information of the prediction and evaluation value calculated offers the recommended project and extracts part 41.

With reference to Figure 15,16 or the like described, this recommended project is extracted part 41 and is extracted the recommended project that will recommend each user according to prediction and evaluation value, project evaluation history and the information relevant with similar users as subsequently.The recommended project is extracted information that part 41 will be used to represent the recommended project extracted and is offered information and present part 42.

Information presents part 42 controls the information relevant with each project is recorded project information retaining part 43 and will record user profile retaining part 44 with each subscriber-related information.In addition, in response to the order of being imported by the importation 21 of user interface part 11 that is used to present project and various information, information presents part 42 and obtains the project and the information of being asked from project information retaining part 43 and user profile retaining part 44, and project and the information obtained are sent to display part 22, thereby control is presented to the user with project and various information.

It should be noted, user interface part 11 and information processing part 12 can be form by individual equipment configuration or be configured to separate equipment.In user interface part 11 and information processing part 12 is under the situation about being formed by the autonomous device configuration, user interface part 11 is to form by disposing such as the such user terminal of personal computer, mobile phone or consumer electronics, and information processing part 12 is by forming such as Web server or the such server configures of application server.In this case, in information handling system 1, a plurality of user interface part 11 are by linking to each other with information processing part 12 such as the such network in internet.Also can dispose configuration information processing section 12 by a plurality of equipment.

Hereinafter, to user interface part 11 be form by user terminal configuration and information processing part 12 be to be described by such situation that server configures forms.

Next, referring to figs. 2 to 16,1 performed processing is described to information handling system.

At first, with reference to the process flow diagram among the figure 2,1 performed project evaluation is obtained to handle and is described to information handling system.For example, when importation 21 inputs of user by user interface part 11 are used to present the order of expectation project, begin this processing, and the information that this order is sent to information processing part 12 is presented part 42.

In step S1, display part 22 presents project.Specifically, information presents part 42 and obtains the relevant information of project of being asked with the user from project information retaining part 43, and this information is sent to the display part 22 of user interface part 11.According to received information, display part 22 shows the relevant information of project of being asked with the user.For example, if the project that the user asked is the music disc collection, show the artist name relevant, album title, title of song, audition sample so and to comment text of this collection of records or the like with this collection of records.

In step S2, project evaluation obtain part 31 obtain the user give present the evaluation of project.Specifically, for example, in audition, purchase, on probation or use after institute presents project, the user is by the evaluation of importation 21 inputs to this project.The evaluation example of being imported this moment comprises evaluation of estimate and the comment text as the numeric representation of the evaluation that gives this project.In addition, the user can directly import evaluation of estimate, perhaps by the user from such as selecting to import evaluation of estimate in the middle of the such selection of " satisfaction ", " satisfied a little ", " neutrality ", " discontented a little " and " being discontented with ".

Replace the user directly to import evaluation of estimate, can on information handling system 1 side, use history or the like to determine evaluation of estimate according to user items.For example, can expect such configuration, promptly such as when the user uses a specific project repeatedly or when user under the situation of TV programme information page or leaf presets the record of project, the user takes the suggestion user that project is spoken highly of such action, the user's evaluation of estimate to this project can be set to automatically high value so.

Importation 21 will be used to represent that the information of the evaluation of the project imported is sent to the project evaluation value and obtains part 31, and this project evaluation is obtained part 31 and obtained the information that is transmitted.

In step S3, project evaluation is obtained the evaluation of 31 pairs of projects of being obtained of part and is carried out record.That is to say that project evaluation is obtained part 31 evaluation of the project obtained is recorded the project evaluation that remains in the historical retaining part 32 in history.After this, the processing end is obtained in this project evaluation.When the processing repetition was obtained in project evaluation, each user of accumulation gave the history of the evaluation of each project in project evaluation history.

Fig. 3 provided exist five user u1 to u5 of information handling system 1,1 couple of five project i1 to i5 of information handling system to handle and according to from as minimum 1 to the situation of the evaluation of estimate of representing each each project as the highest 5 grade the example of the project evaluation history relevant with evaluation of estimate.Value representation in every row of the project evaluation history among Fig. 3 and this are listed as the evaluation that corresponding user gives to be listed as with this corresponding project.For example, in Fig. 3, the evaluation of estimate that user u1 gives project i2 is 5, and the evaluation of estimate that user u5 gives project i5 is 3.Each blank column in the project evaluation history represents that being listed as corresponding user with this does not make evaluation to being listed as corresponding project with this.

Hereinafter, be that processing under the situation about being kept by historical retaining part 32 specifically describes to the project evaluation history among Fig. 3.

Next, with reference to the process flow diagram among the figure 4,1 performed item characteristic computing is described to information handling system.

In step S21, project statistical computation part 33 is obtained the project evaluation history that historical retaining part 32 is kept.

In step S22, project statistical computation part 33 is come the computational item statistics according to this project evaluation history.Project statistics comprises at least three statistics, is used to promptly to represent that the evaluation number N i of the number of the evaluation that given, the evaluation mean value avg (Ri) and being used to that is used to represent the mean value of evaluation of estimate represent the evaluation variance var (Ri) of the variance of evaluation of estimate.

Estimate number N i and represent that the user organizes the interested degree of this project.Usually, the evaluation number that gives each project presents so-called long-tail (long tail) trend, makes that the very evaluation of big figure is the center with the popular project of suitable minority purpose, and gives the seldom evaluation of number to the project of other broad range.Therefore, replace estimating number N i, but the logarithm log Ni of in-service evaluation number N i.Hereinafter, the logarithm log Ni that also will estimate number N i is called main index (majorness index) Mi.

Estimating mean value avg (Ri) is bad standard with acting on definite project of paying close attention to.

The evaluation of estimating the central project of being paid close attention to of variance var (Ri) expression user changes.

Fig. 5 has provided the project statistics of being calculated according to the project evaluation history among Fig. 3.Among Fig. 5 second row shows the evaluation number N i and the main index M i (number in the bracket) of each each project, the third line shows the evaluation mean value avg (Ri) of each each project, and fourth line shows the evaluation variance var (Ri) of each each project.For example, in Fig. 5, for project i1, estimating number N 1 is 2, and main index M 1 is 0.69, and estimating mean value avg (R1) is 4.5, and evaluation variance var (R1) is 0.25.

The such processing of project statistics that project statistical computation part 33 repeats to select the project (being called the concern project hereinafter) that will pay close attention to and calculates the concern project becomes the concern project until all items, thereby calculates the project statistics of each project.Project statistical computation part 33 will be used to represent that the information of the project statistics of each project of being calculated offers item types determining section 34.

In step S23, item types determining section 34 obtains the rank or the set of each project according to the project statistics.Specifically, for example, item types determining section 34 comes project is carried out rank according to each statistics (estimate number N i, estimate mean value avg (Ri) and estimate variance var (Ri)) that is included within the project statistics.

Fig. 6 has provided the rank of project when carrying out rank according to the statistics of the project among Fig. 5.Among Fig. 6 second row has provided the rank Pni when arranging project according to the incremental order of estimating number N i, the third line has provided the rank Pai when arranging project according to the incremental order of estimating mean value avg (Ri), and fourth line has provided the rank Pvi when arranging project according to the incremental order of estimating variance var (Ri).For example, in Fig. 6, the rank according to estimating number of project i1 is 1, is 5 according to the rank Pa1 that estimates mean value, and is 3 according to the rank Pv1 that estimates variance.

Alternatively, for example, item types determining section 34 comes project is divided into groups according to each statistics (estimate number N i, estimate mean value avg (Ri) and estimate variance var (Ri)) that is included within the project statistics by using any threshold.For example, item types determining section 34 becomes to estimate number N i by in-service evaluation number N i with group items and is equal to or greater than the main project S set mj of threshold value and estimates the less important project set Smn of number N i less than threshold value, become evaluation mean value avg (Ri) to be equal to or greater than the high praise project set Sah and the low evaluation project set Sal of evaluation mean value avg (Ri) of threshold value group items by in-service evaluation mean value avg (Ri), perhaps group items one-tenth is estimated the evaluation variation project set Svh greatly and the evaluation variation little project set Svl of evaluation variance var (Ri) less than threshold value that variance var (Ri) is equal to or greater than threshold value by in-service evaluation variance var (Ri) less than threshold value.

In step S24, item types determining section 34 type of identifying project.For example, under the situation of project implementation rank, item types determining section 34 is by carrying out the item types that appropriate combination is determined each each project to the rank that is obtained in step S23.For example, item types determining section 34 is determined the masterpiece index M Pi of each each project according to following equation (1).

Masterpiece index M Pi=presses the rank Pvi that estimates variance by the rank Pni+ that estimates number by the rank Pai-that estimates mean value ... (1)

That is to say, when estimate number become big, when estimating mean value and uprising and estimates variance and diminish, masterpiece index M Pi becomes greatly.Therefore, the project with high masterpiece index M Pi receives the high mean value evaluation from a large amount of people.Item types determining section 34 is defined as " masterpiece " with the item types that for example its masterpiece index M Pi is equal to or higher than the project of predetermined threshold.

In addition, for example, item types determining section 34 is determined the hiding masterpiece index SMPi of each each project according to following equation (2).

Hide masterpiece index SMPi=-by the rank Pai of the rank Pni+ that estimates number ... (2) by evaluation mean value

That is to say, when estimating that number diminishes and when estimating mean value and uprising, hiding masterpiece index SMPi and become greatly.Therefore, the project with the high masterpiece index SMPi of hiding receives the high mean value evaluation from a small amount of people.Item types determining section 34 for example is defined as the item types that its hiding masterpiece index SMPi is equal to or higher than the project of predetermined threshold " hiding masterpiece ".

Fig. 7 has provided based on the masterpiece index M Pi of each each project of the rank of the project among Fig. 6 and hiding masterpiece index SMPi.Among Fig. 7 second row has provided the masterpiece index M Pi of each each project, and the third line has provided the hiding masterpiece index SMPi of each each project.For example, in Fig. 7, the masterpiece index M P1 of project i1 is 3, and its hiding masterpiece index SMP1 is 4.

In addition, for example, under the situation of project implementation grouping, item types determining section 34 is determined the item types of each project by the combination of the set under each project in step S23.For example, have big evaluation number N i, high evaluation mean value avg (Ri) and little evaluation variance var (Ri) because be included in productive set the project within the Smj ∩ Sah ∩ Svl of closing, so item types determining section 34 is defined as " masterpiece " with the item types of this project.In addition, have little evaluation number N i and high evaluation mean value avg (Ri), so item types determining section 34 is defined as " hiding masterpiece " with the item types of this project because be included in the project that productive set closes within the Smn ∩ Sah.

The such processing of item types that item types determining section 34 repeats to select a concern project and determines this concern project becomes the concern project until all items, thereby determines the item types of each project.Item types determining section 34 will be used to represent that the information of the item types of determined each project offers information and presents part 42.Information presents part 42 item types of determined each project is added on the information of each project that project information retaining part 43 kept.

In step S25, information presents part 42 item types is presented to the user.For example, when by the processing identical with step S1 among Fig. 2 the information of project being presented to the user, information presents the information that part 42 also will be used to represent the item types of this project and is sent to display part 22.This display part 22 shows the item types (for example " masterpiece ", " hiding masterpiece " or the like) of this project and the information of the project that the user asked.

In this manner, the user who gives each project by effective use estimates, and can suitably determine the item types of each each project and determined item types is presented to the user.Therefore, the user can learn the evaluation trend that gives each each project.

Next, with reference to the process flow diagram among the figure 8, information handling system 1 performed similar terms is extracted processing be described.

In step S41, as in the processing of the step S21 among Fig. 4, project statistical computation part 33 is obtained project evaluation history.After this, in step S42, as in the processing of the step S22 among Fig. 4, project statistical computation part 33 computational items statistics, and will be used to represent that the information of the project statistics calculated offers project index of similarity calculating section 35.

In step S43, project index of similarity calculating section 35 computational item index of similarity.For example, project index of similarity calculating section 35 by use with regard to the difference between the main index M j of the main index M i of project i and project j monotonically increasing function come between computational item i and the project j index of similarity Sim (i, j).

Sim (i, j)=1/ (| Mi-Mj|+ ε) (ε is positive constant) ... (3)

That is to say poor when the main index between the project | when Mi-Mj| diminished, (i j) became big to the project index of similarity Sim that is obtained from equation (3), and this represents that these two projects are similar each other.

Fig. 9 has provided that (1, j), wherein ε is set to equal 0.01 by the project i1 that utilizes equation (3) and calculated and the index of similarity Sim between each other each project according to the main index M i among Fig. 5.For example, in Fig. 9, project index of similarity Sim (1 between project i1 and the project i2,2) be 2.41, the project index of similarity Sim (1,3) between project i1 and the project i3 is 1.08, project index of similarity Sim (1 between project i1 and the project i4,4) be 1.42, and the project index of similarity Sim (1,5) between project i1 and the project i5 is 2.41.

Also can be defined as vi=(Mi by vector with project i, avg (Ri), var (Ri)) and with the vector of project j be defined as vj=(Mj, avg (Rj), var (Rj)), and the function (for example inverse of Euclidean distance) of utilization monotone decreasing with regard to the Euclidean distance between vector vi and the vector vj, come computational item index of similarity Sim (i, j), perhaps the cosine index of similarity between compute vectors vi and the vector vj with as project index of similarity Sim (i, j).In this case, the distribution trend of value that has constituted each element (main index, estimate mean value and estimate variance) of vector vi and vj differs from one another.Therefore, for each element, for make mean value become 0 and variance become 1 and the value of regularization is set to the value of each element of vector vi and vj.

Project index of similarity calculating section 35 repeats following processing, promptly select a concern project and calculate project index of similarity Sim (i between this concern project and another project, j), the concern project is changed, until calculated the project index of similarity Sim in the middle of all items (i, j).(i j) offers similar terms and extracts part 36 project index of similarity calculating section 35 with the project index of similarity Sim that calculated.

By not only utilizing project statistics but also utilize the information relevant with each each project, can obtain new project index of similarity Sim ' (i, j).For example, if project is a document, can expect such configuration so, promptly utilize the frequency of occurrences of each speech in each project usually to create the speech vector as unit, and by utilizing the cosine distance C os (i between the speech vector, j) and based on the above-mentioned project index of similarity Sim of project statistics (i, j) and by following equation (4) calculate new project index of similarity Sim ' (i, j).

Sim′(i，j)＝Cos(i，j)+Sim(i，j) ...(4)

In step S44, similar terms extracts part 36 and extracts similar terms.For example, similar terms extracts part 36 and repeats following processing, promptly select a concern project, and extract its project index of similarity Sim (i for this concern project, j) for example be equal to or higher than the similar terms of the project of predetermined threshold as this concern project, become the concern project until all items, thereby extract the similar terms of each project.

Alternatively, similar terms extracts part 36 and repeats following processing, promptly extract and work as according to the project index of similarity Sim (i for the concern project, top n project when the order of successively decreasing j) comes project sorted is as the similar terms of the project of concern, become the concern project until all items, thereby extract the similar terms of each project.For example, if N=2 under the situation of the project index of similarity in Fig. 9 extracts so and has the highest two index of similarity Sim (1, project i2 j) and project i5 are as the similar terms of project i1.

Similar terms extracts information that part 36 will be used to represent the similar terms of each project of being extracted and offers information and present part 42.Information presents part 42 information of the similar terms of each project of being extracted is added on the information of each project that project information retaining part 43 kept.

In step S45, information presents part 42 similar terms is presented to the user.For example, when by the processing identical with step S1 among above-mentioned Fig. 2 the information of project being presented to the user, information presents the information that part 42 also will be used to represent the similar terms of this project and is sent to display part 22.Information that display part 22 demonstrations are relevant with the project that the user is asked and the information relevant with the similar terms of this project.

In this manner, the user who gives each project by effective use estimates, can suitably extract have similar evaluation trend project to present to the user.

Though top description at for calculate for each project its for another project the project index of similarity and extract the situation of similar terms, can be only to this processing of the essential project implementation of for example institute's request items and so on.In addition, by utilizing various conditions to limit the scope of the similar terms that will extract at (for example style, issuing date or the like).

Next, with reference to the process flow diagram among Figure 10,1 performed user characteristics computing is described to information handling system.

In step S61, as in the processing of the step S21 among Fig. 4, project statistical computation part 33 is obtained project evaluation history.After this, in step S62, as in the processing of the step S22 among Fig. 4, project statistical computation part 33 computational items statistics, and will be used to represent that the information of the project statistics calculated offers user's statistical computation part 37.

In step S63, user's statistical computation part 37 is calculated user's statistics.Now, the accounting example that is included within user's statistics is described.

For example, be included in the mean value avg_u (Mi) of the main index M i that pays close attention to the project within the project set Cu that user u estimated and variance var_u (Mi) has given evaluation to which intermediate item as user u index.Especially, mainly the mean value avg_u (Mi) of index M i expression gives the mean value of the evaluation number N i of user's project that u is estimated.If should the value very big, so as can be known user u tend to popular project interested, and if should the value very little, user u tends to unfashionable project interested so as can be known.That is to say, can think faddict's rank of mean value avg_u (Mi) expression user of main index M i.Therefore, hereinafter, also the mean value avg_u (Mi) with main index M i is called faddict's index M Hu.In addition, hereinafter, the variance var_u (Mi) of main index M i is called main index variance var_u (Mi).

Figure 11 has provided each each user's who is calculated according to the project statistics among the history of the project evaluation among Fig. 3 and Fig. 5 faddict's index M Hu and main index variance var_u (Mi).Among Figure 11 second row has provided the main index M i of project i1 to i5, second to six row in the 3rd to the 7th row have provided the main index M i of the project that user u1 to u5 estimated, and the 7th row in the 3rd to the 7th row have provided faddict's index M hu of user u1 to u5, and the 8th row in the 3rd to the 7th row have provided the main index variance var_u (Mi) of user u1 to u5.For example, in Figure 11, faddict's index M H1 of user u1 is 1.27, and main index variance var_1 (Mi) is 0.058.

In addition, user u is included in the related coefficient Cor (Rui between the main index M i of the evaluation of estimate Rui of the project of set within the Cu and this project, Mi) expression user u evaluation of estimate Rui that gives user's project that u is estimated and the mean value of estimating number N i (or rather, estimate the mean value of the logarithm of number N i) between correlativity, and tend to the index spoken highly of to any intermediate item as user u.For example, if related coefficient Cor (Rui, Mi) very big, this means that so user u tends to speak highly of to the interested project of a lot of people.Therefore, can think that user u has main orientation or similar follower's (follower) feature.

In addition, the related coefficient Cor (Rui, avg (Ri)) between the evaluation mean value avg (Ri) of the user u evaluation of estimate Rui that is included in the project of set within the Cu and this project is used as whether user u is the index of average user.For example,, can think that so user u is very common, that is to say that user u has average value and sees if related coefficient Cor (Rui, avg (Ri)) is very big.

User's statistical computation part 37 repeats the user's (be called hereinafter and pay close attention to the user) who selects will pay close attention to and the user who calculates this concern user added up such processing, become until all users and pay close attention to the user, thereby the user who calculates each user adds up.

(Rui Mi), perhaps can only calculate the necessary one or more values in these values, adds up as the user can to calculate whole above-mentioned faddict's index M Hu, main index variance var_u (Mi) and related coefficient Cor.

In step S64, user's statistical computation part 37 is calculated user's relative statistic.Now, the example that is included in the relative statistic within user's relative statistic is described.

For example, pay close attention to the faddict rank of user for all users as the faddict's index M hu that pays close attention to the user with the relative faddict's index M Hu-avg (MHu) that departs from of the mean value avg (MHu) of faddict's index of all users.For example, can think that the user with big relative faddict's index M Hu-avg (MHu) has especially strong faddism feature in the middle of all users.

Figure 12 has provided relative faddict's index of each each user who is calculated according to the faddict's index among Figure 11.For example, in Figure 12, relative faddict's index M H1-avg (MHu) of user u1 is-0.004.

User's statistical computation part 37 repeats to select a such processing of user's relative statistic of paying close attention to the user and calculating this concerns user, becomes the concern user until all users, thereby calculates user's relative statistic of each user.After this, user's statistical computation part 37 will be used to represent that the information of user's statistics of each user and user's relative statistic offers information and presents part 42.Information presents part 42 user that obtained statistics and user's relative statistic is added on each user's that user profile retaining part 44 kept the information.

In step S65, information presents part 42 and according to user's statistics and user's relative statistic user's feature is presented to the user.For example, when being used to present the order of the information relevant with user A by importation 21 inputs, information presents the feature of part 42 according to user statistics and user's relative statistic acquisition user A, and the feature of the user A that obtained is added on the information of user A and with this information and be sent to display part 22.The information of the feature of display part 22 explicit user A and the requesting users A of institute.For example, on the SNS of overview that shows user A or the like (social networking service) " my page ", show such as " faddict's index of user A: ★ ★ ★ ★ ☆ " according to faddict's index M Hu or relative faddict's index M Hu-avg (MHu).

In this manner, the user who gives each project by effective use estimates, and the feature that can obtain each user exactly is to present to the user.

Next, with reference to the process flow diagram among Figure 13, information handling system 1 performed similar users is extracted processing be described.

In step S81, as in the processing of the step S21 among Fig. 4, project statistical computation part 33 is obtained project evaluation history.After this, in step S82, as in the processing of the step S22 among Fig. 4, project statistical computation part 33 computational items statistics, and will be used to represent that the information of the project statistics calculated offers user's statistical computation part 37.

In step S83, as in the processing of the step S63 among Figure 10, user's statistical computation part 37 is calculated users' statistics, and will be used to represent that the information of user's statistics of being calculated offers user's index of similarity calculating section 38.

In step S84, user's index of similarity calculating section 38 is added up according to the user and is calculated user's index of similarity.For example, the main index M i that is included in the project within the project set that each user estimates by supposition is in normal distribution, KL between user's index of similarity calculating section 38 calculates the main index M i be included in the project within the project set Cu that user u estimated from following equation (5) distribution and the distribution of the main index M i of the project set Cv that user v is estimated is apart from (Kullback-Leibler divergence, Kullback-Lai Baile divergence), as between the user between user u and the user v apart from D (u, v).

$D(u,v)=\frac{1}{2}(\log \left(\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="v"\; filenames="A200810178965E00631.png,A200810178965E00651.png"\; state="\{\{state\}\}">v</figure-callout>}}^2}{{\mathrm{\&sigma;}}_u^2}\right)+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="A200810178965E00631.png,A200810178965E00651.png"\; state="\{\{state\}\}">u</figure-callout>}}^2}{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="v"\; filenames="A200810178965E00631.png,A200810178965E00651.png"\; state="\{\{state\}\}">v</figure-callout>}}^2}+\frac{{({\mathrm{\&mu;}}_v-{\mathrm{\&mu;}}_u)}^2}{{\mathrm{\&sigma;}}_v^2}-1)\&CenterDot;\&CenterDot;\&CenterDot;\left(5\right)$

In equation (5), μ _uThe mean value avg_u (Mi) (that is to say faddict's index M Hu) of the main index M i of the project among the project set Cu that expression user u is estimated, σ _u ²The main index variance var_u (Mi) of the project among the expression set Cu, μ _vThe mean value avg_v (Mi) (that is to say faddict's index M Hv) of the main index M i of the project among the project set Cv that expression user v is estimated, and σ _v ²The main index variance var_v (Mi) of the project among the expression set Cv.

Because KL distance does not become symmetry with regard to u and v, therefore can obtain (D (u, v)+D (v, u))/2 is as distance between the user between user u and the user v.

After this, user's index of similarity calculating section 38 by utilize as shown in following equation (6) with respect between the user apart from D (u, v) the function of monotone decreasing calculate user's index of similarity SimU between user u and the user v (u, v).

SimU(u，v)＝1-D(u，v) ...(6)

Figure 14 provided according to the user among Figure 11 statistics pass through between user u3 that above-mentioned equation (5) and equation (6) calculated and the user between other users apart from D (u, v) with user's index of similarity SimU (u, v).For example, in Figure 14, be 0.25 apart from D (3,1) between the user between user u3 and the user u1, and user's index of similarity SimU (3,1) is 0.75.

User's index of similarity calculating section 38 repeats following processing, promptly select one pay close attention to the user and calculate this concern user and another user between the user between apart from D (u, v) with user's index of similarity SimU (u, v), change simultaneously and pay close attention to the user, until calculated user distance D in the middle of the every other user (u, v) with user's index of similarity SimU (u, v).User's index of similarity calculating section 38 will be used to represent that (u, information v) offers similar users and extracts part 39 for user's index of similarity SimU of being calculated.

In step S85, similar users is extracted part 39 and is extracted similar users.For example, similar users is extracted part 39 and is repeated following processing, promptly select one to pay close attention to the user and extract its user's index of similarity SimU (u for this concern user, the user who v) equals or be higher than predetermined threshold for example is with the similar users as this concern user, become the concern user until all users, thereby extract each user's similar users.Alternatively, similar users is extracted part 39 and is repeated following processing, promptly extracts (u, descending v) and the top n user that sorts according to the user's index of similarity SimU for paying close attention to the user, become the concern user until all users, thereby extract each user's similar users.

Similar users is extracted information that part 39 will be used to represent each user's of being extracted similar users and is offered information and present part 42.Information presents part 42 information of each user's of being extracted similar users is added on each user's that user profile retaining part 44 kept the information.

In step S86, information presents part 42 similar users is presented to the user.For example, when being used to present the order of the information relevant with user A by importation 21 input, information presents the information that part 42 will be used to represent the similar users of user A and is sent to display part 22 with the information of user A.The information of the similar users of display part 22 explicit user A and the requesting users A of institute.For example, go up at the SNS of overview that shows user A or the like (social networking service) " my page " and show that the similar users tabulation is with as " user similar to user A ".

In this manner, the user who gives each project by effective use estimates, can suitably extract the user that tends to give similar evaluation to present to the user, that is to say to project, can suitably extract have similar values and preference the user to present to the user.

Though above description is at calculating it with respect to another purpose project index of similarity and extract the such situation of similar terms for each project, can be only to for example the so necessary user of the user that asks carry out this processing.In addition, can limit the scope of the similar users that will extract by utilizing various conditions (for example sex, age and address).

Next, with reference to the process flow diagram among Figure 15,1 performed project recommendation is handled and is described to information handling system.

In step S101, identical with the processing of step S21 among Fig. 4, project statistical computation part 33 is obtained project evaluation history.After this, identical with the processing of step S22 among Fig. 4 in step S102, project statistical computation part 33 computational items statistics, and will be used to represent that the information of the project statistics calculated offers user's statistical computation part 37.

In step S103, identical with the processing of step S63 among Figure 10, user's statistical computation part 37 is calculated users' statistics, and will be used to represent that the information of user's statistics of being calculated offers user's index of similarity calculating section 38.

In step S104, identical with the processing of step S84 among Figure 10, user's index of similarity calculating section 38 calculates user's index of similarity, and will be used to represent that the information of user's index of similarity of being calculated offers prediction and evaluation value calculating section 40.

In step S105, prediction and evaluation value calculating section 40 calculates the prediction and evaluation value.For example, (u v) calculates the prediction and evaluation value Rui ' of the user u for the unvalued project i of user u according to following equation (7) by utilizing user's index of similarity SimU.

$R_{\mathrm{ui}}\&prime;=\mathrm{avg}\_R_u+\frac{\underset{v}{\mathrm{\&Sigma;}}(R_{\mathrm{vi}}-\mathrm{avg}\_R_v)\mathrm{SimU}(u,v)}{\underset{v}{\mathrm{\&Sigma;}}\left|\mathrm{SimU}(u,v)\right|}\&CenterDot;\&CenterDot;\&CenterDot;\left(7\right)$

In equation (7), ave_Ru represents that user u is included in the mean value of the evaluation of estimate of the project within the project set Cu that user u estimated, avg_Rv represents that user v is included in the mean value of the evaluation of estimate of the project within the project set Cv that user v estimated, and Rvi represents that user v gives the evaluation of estimate of project i.In equation (7), do not use the user's data of project i not being made evaluation.

According to equation (7), big weight assignment is given the big index of similarity SimU (u that has for user u, evaluation of user value Rvi v), and little weight assignment given little index of similarity SimU (u, the evaluation of user value Rvi v) have for user u.Therefore, (the evaluation of estimate Rvi that u, user v) give project i further greatly is reflected on the prediction and evaluation value Rui ' to have big index of similarity SimU for user u.

It should be noted, at above-mentioned P.Resnick, N.Iacovou, M.Suchak, P.Bergstrom and J.RIedl " GroupLens:Open Architecture forCollaborative FilterIng of Netnews " (Conference on Computer SupportedCooperative Work, the 175-186 page or leaf, 1994) in the disclosed example, replace the SimU (u in the equation (7), v), used Pearson correlation coefficient with regard to the evaluation of estimate between user u and the user v.

Predictor calculation part 40 repeats to select a concern user, selects a concern project and calculate the prediction and evaluation value Rui ' such processing of this concern user to this concern project in the middle of the unvalued project of this concern user, become the concern project until the unvalued all items of this concern user, and become the concern user until all users, thereby calculate the prediction and evaluation value of each user unvalued each project.Predictor calculation part 40 will be used to represent that the information of prediction and evaluation value Rui ' offers recommended project extraction part 41.

In step S106, the recommended project is extracted part 41 and is extracted the recommended project.For example, the recommended project extract part 41 repeat to select one pay close attention to the user and extract that this concern user equals its prediction and evaluation value Rui ' or the project that is higher than predetermined threshold with as the such processing of the recommended project, become the concern user until all users, thereby extract each user's the recommended project.In addition, for example, the recommended project is extracted part 41 and is repeated to select one to pay close attention to the user and extract by the top n project of the descending sort of this concern user's prediction and evaluation value Rui ' with as the such processing of the recommended project, become the concern user until all users, thereby extract each user's the recommended project.

The recommended project is extracted information that part 41 will be used to represent each user's the recommended project and is offered information and present part 42.Information presents part 42 information of the recommended project extracted is added on each user's that user profile retaining part 44 kept the information.

In step S107, information presents part 42 institute's recommended project is presented to the user.For example, as required, information presents part 42 and will be used to represent be sent to display part 22 as the information of the possessory user's of user interface part 11 the recommended project.This display part 22 shows recommended project tabulation.

In this manner, the user who gives each project by effective use estimates, and can recommend suitable project to each user.

Next, with reference to the process flow diagram among Figure 16, second embodiment that project recommendation is handled is described.

In step S121, identical with the processing of step S21 among Fig. 4, project statistical computation part 33 is obtained project evaluation history.After this, identical with the processing of step S22 among Fig. 4 in step S122, project statistical computation part 33 computational items statistics and will being used to represents that the information of the project statistics calculated offers user's statistical computation part 37.

In step S123, identical with the processing of step S63 among Figure 10, user's statistical computation part 37 is calculated users' statistics and the information of user's statistics that will be used to represent calculated offers user's index of similarity calculating section 38.

In step S124, identical with the processing of step S84 among Figure 10, user's index of similarity calculating section 38 calculates user's index of similarity and the information of user's index of similarity that will be used to represent calculated offers similar users and extracts part 39.

In step S125, identical with the processing of step S85 among Figure 13, similar users extracts that part 39 is extracted similar users and the information of the similar users that will be used to represent extracted offers the recommended project and extracts part 41.

In step S126, the recommended project is extracted part 41 and is extracted the recommended project.Specifically, the recommended project is extracted part 41 and is obtained the project evaluation history that historical retaining part 32 is kept.The recommended project is extracted part 41 and is repeated following processing, promptly select one pay close attention to the user and extraction has been given high evaluation value by this concern user's similar users in the middle of the unvalued project of this concerns user project with as the recommended project, become the concern user until all users, thereby extract each user's the recommended project.For example, the number or the ratio of similar users that extract the mean value of the evaluation of estimate that similar users gave or project that mxm. is equal to, or greater than predetermined threshold, has given to be equal to or greater than the evaluation of estimate of predetermined threshold is equal to or greater than project of predetermined threshold or the like, with the recommended project as the concern user.

For example, according to the user's index of similarity SimU (u among Figure 14, v) select under the situation of user u5 as the similar users of user u3, if being used to extract the threshold value of the evaluation of estimate of the recommended project is 3, so according to the project evaluation history among Fig. 3, the evaluation of estimate of selecting user u5 to give it in the middle of the unvalued project of user u3 is equal to, or greater than 3 project i5 with the recommended project as user u3.

The recommended project is extracted information that part 41 will be used to represent each user's the recommended project and is offered information and present part 42.Information presents part 42 information of the recommended project extracted is added on each user's that user profile retaining part 44 kept the information.

In step S127, identical with the processing of step S107 among Figure 15, the recommended project is presented to the user.

In this manner, the user who gives each project by effective use estimates, and suitable project can be recommended each user.

As mentioned above, the user who gives each project by effective use estimates, and can obtain to be not easy from the social status of each project that the description (metadata) of project is understood or each user's social status.In addition, can reflect the user's of other similar types preference, thereby can recommend and the better project of mating of user preference.

More than describe at information processing part 12 and collect the such situation of evaluation that each user gives each project.Yet, can expect that also for example information processing part 12 obtains the collected project evaluation of another equipment and carry out the such embodiment of above-mentioned processing.

Now, with reference to Figure 17 and 18, another example of item types is described.Figure 17 has gathered to be used to calculate to be used for identifying project the form of formula of each index of type.Figure 18 is the evaluation mean value that has gathered project, the form of estimating the relation between variance, evaluation number and each item types.

As mentioned above, obtain the masterpiece index by " pressing the rank Pvi that estimates variance by the rank Pai-that estimates mean value " by the rank Pni+ that estimates number.When estimate number become big, when estimating mean value and uprising and estimates variance and diminish, the masterpiece index becomes greatly.That is to say that the project with high masterpiece index is the project that receives from the high evaluation of a large number of users.

Not only can be but also can obtain to hide the masterpiece index by " by the rank Pni+ that estimates number by the rank Pai-that estimates mean value by the rank Pvi that estimates variance " by " by the rank Pni+ that estimates number by the rank Pai that estimates mean value ".Under latter instance, when estimating number and diminish, estimate mean value and uprise and estimates variance and diminish, hide the change of masterpiece index greatly.That is to say that having the high project of hiding the masterpiece index is the project that receives from a small amount of people's high evaluation.

Obtain dispute works (controversial piece) index by " pressing the rank Pvi that estimates variance by the rank Pai+ that estimates mean value " by the rank Pni+ that estimates number.When estimate number become big, when estimating mean value and uprising and estimates variance and become big, dispute works index becomes greatly.That is to say, can think that the project with high dispute works index is the high evaluation project that still its evaluation greatly changes from the user to the user that receives from many people, but promptly talked about the project of mixing evaluation that receives in a large number.Item types determining section 34 is equal to, or greater than its dispute works index for example, and the item types of the project of predetermined threshold is defined as " dispute works ".

Obtain the enthusiast by " pressing the rank Pvi that estimates variance by the rank Pai+ that estimates mean value " and appeal index by the rank Pni+ that estimates number.When estimating number and diminish, estimate mean value and uprise and estimates variance and become big, the enthusiast appeals the index change greatly.That is to say, can think that having project that high enthusiast appeals index is to receive high evaluation but it estimates the project that greatly changes from the user to the user from a small amount of people, be i.e. some projects of liking.Item types determining section 34 appeals for example to be equal to, or greater than by index with its enthusiast, and the item types of the project of predetermined threshold is defined as " enthusiast's appealing ".

Obtain useless works index by " pressing the rank Pvi that estimates variance by the rank Pai-that estimates mean value " by the rank Pni-that estimates number.Become big, estimate the mean value step-down and estimate variance when diminishing when estimating number, it is big that useless works index becomes.That is to say, can think that the project with high useless works index is to receive the project that harmonic(-)mean is estimated from a large amount of people, but promptly talked about project in a large number with bad quality.Item types determining section 34 is equal to, or greater than its useless works index for example, and the item types of the project of predetermined threshold is defined as " useless works ".

Obtain to merit no attention (unworthy of attention) index by " pressing the rank Pvi that estimates variance by the rank Pai-that estimates mean value " by the rank Pni-that estimates number.Diminish, estimate the mean value step-down and estimate variance when diminishing when estimating number, the index that merits no attention becomes greatly.That is to say, can think that the project with the high index that merits no attention is to receive the project that harmonic(-)mean is estimated from a small amount of people, be i.e. the project noted of nobody almost.Item types determining section 34 is equal to, or greater than its index that merits no attention for example, and the item types of the project of predetermined threshold is defined as " meriting no attention ".

By "+press the rank Pvi that estimates variance by the rank Pai+ that estimates mean value by the rank Pni-of evaluation number " a large amount of item (the mass produced piece) indexes that produce of acquisition.Diminish, estimate the mean value step-down and estimate variance when diminishing when estimating number, a large amount of index that produce become greatly.That is to say, can think that but the project with high a large amount of production works indexes is to receive the harmonic(-)mean evaluation the project that its evaluation greatly changes from the user to the user from a large amount of people, but promptly talked about project in a large number with imperfect quality.Item types determining section 34 for example is equal to, or greater than that the item types of the project of predetermined threshold is defined as " a large amount of produce " with its a large amount of indexes that produce.

Obtain rough works (crude piece) index by " pressing the rank Pvi that estimates variance by the rank Pai+ that estimates mean value " by the rank Pni-that estimates number.Diminish, estimate the mean value step-down and estimate variance when becoming big when estimating number, rough works index becomes greatly.That is to say, can think that the project with high rough works index is to receive the harmonic(-)mean evaluation from a small amount of people it estimates the project that greatly changes from the user to the user, although but the i.e. project liked of its less important some people.Item types determining section 34 is equal to, or greater than its rough works index for example, and the item types of the project of predetermined threshold is defined as " rough works ".

In addition, for example, item types determining section 34 is defined as " mainly " with the item types that its main index is equal to, or greater than the project of predetermined threshold A, and the item types that its main index is lower than the project of this threshold value A is defined as " less important ".

In the following description, with the evaluation of estimate Rui of user u and the related coefficient Cor (Rui between the main index M i, Mi) be called " main orientation index ", and the evaluation of estimate Rui of user u and the related coefficient Cor (Rui, avg (Ri)) that estimates between the mean value avg (Ri) are called " general index number ".

Next, referring to figures 19 through 54, the second embodiment of the present invention is described.

Figure 19 is the block scheme that has provided according to the information handling system of second embodiment of the invention.Information handling system 101 among Figure 19 comprises user interface part 111 and information processing part 112.User interface part 111 comprises importation 121 and display part 122.Information processing part 112 comprises that project evaluation obtains part 131, historical retaining part 132, project statistical computation part 133, item types determining section 134, project index of similarity calculating section 135, similar terms extracts part 136, user's statistical computation part 137, user's index of similarity calculating section 138, similar users is extracted part 139, prediction and evaluation value calculating section 140, the recommended project is extracted part 141, information presents part 142, project information retaining part 143, user profile retaining part 144, user's cluster produces part 145, project cluster generating unit divides 146, and present regular retaining part 147.

In this accompanying drawing, be to represent with corresponding part among Fig. 1 by reference number identical among its final two digits and Fig. 1, and omit to the description of the corresponding part of similar processing to avoid repetition.

As described with reference to Figure 34 or the like subsequently, the history of project information that project statistical computation part 133 is kept according to historical retaining part 132 is calculated the project statistics that is used to represent give the evaluation trend of each each project.Project statistical computation part 133 will be used to as required to represent that the information of the project statistics calculated offers item types determining section 134, project index of similarity calculating section 135, user's statistical computation part 137 and information and presents part 142.

As described with reference to Figure 20 or the like subsequently, according to project evaluation history that historical retaining part 132 kept, present part 142 and the 146 project cluster informations that provide are provided for the project information of being obtained from project information retaining part 143, the project statistics that project statistical computation part 133 is provided, user's cluster information and the project cluster generating unit that user's cluster produces part 145 and provided by information, user's statistical computation part 137 is calculated user's statistics of the feature that is used to represent each user according to the evaluation trend that gives each project.User's statistical computation part 137 will be used to as required to represent that the information of user's statistics of being calculated offers user's index of similarity calculating section 138 and information presents part 142.

As described with reference to Figure 48 or the like subsequently, the recommended project in Fig. 1 is extracted the processing of part 41, and the recommended project is extracted part 141 and also extracted the project that will present to each user according to presenting the project information that part 142 obtained from project information retaining part 143 by information and presenting part 142 by information from the user profile that user profile retaining part 144 is obtained.The recommended project is extracted information that part 141 will be used to represent the project extracted and is offered information and present part 142.

As described with reference to Figure 45 or the like subsequently, information in Fig. 1 presents the processing of part 42, information present part 142 also according to presenting of presenting that regular retaining part 147 kept project information that rule, project information retaining part 143 kept and the user profile that kept of user profile retaining part 144 control by display part 12 and present the information relevant with project.

As described with reference to Figure 21 or the like subsequently, user's cluster produces part 145 and makes user's cluster according to the project evaluation history that historical retaining part 132 is kept by utilizing preordering method to carry out.User's cluster produces part 145 the user cluster information relevant with the user's cluster that produces as clustered result is offered user's statistical computation part 137.

As described with reference to Figure 23 or the like subsequently, project cluster generating unit is divided the 146 project evaluation history that kept according to historical retaining part 132, is come the project implementation cluster by utilizing preordering method.Project cluster generating unit divides 146 the project cluster information relevant with the project cluster as clustered result that is produced offered user's statistical computation part 137.

Presenting regular retaining part 147 obtains and keeps presenting rule.Presented rule predetermining when the rule that will when the information relevant with project of importation 121 inputs outside input or by user interface part 111 is presented to the user, will follow.

Next, referring to figures 20 through 54,101 performed processing are described to information handling system.

The same with information handling system 1, but the project evaluation in information handling system 101 execution graphs 2 is obtained that item characteristic computing among processing, Fig. 4, the similar terms among Fig. 8 are extracted and handled, user characteristics computing, the similar users among Figure 13 among Figure 10 are extracted and handled, the project recommendation among Figure 15 is handled and Figure 16 in project recommendation handle.Omission to the description of these processing to avoid repetition.

At first, referring to figures 20 through 42, the processing that information handling system 101 is obtained user and item characteristic is described.

At first, with reference to the process flow diagram among Figure 20, user characteristics (reputation orientation index) computing that is used to calculate the reputation orientation index of representing class user statistics is described.

In step S201, identical with the processing of step S21 among Figure 21, project statistical computation part 133 is obtained project evaluation history.Hereinafter, the processing under the situation of obtaining the project evaluation history among Fig. 3 is specifically described.

In step S202, project statistical computation part 133 is according to the historical evaluation mean value that calculates each project of project evaluation.Therefore, the evaluation mean value of each project shown in the third line of calculating chart 5.Project statistical computation part 133 will be used to represent that the information of the evaluation mean value of each project of being calculated offers user's statistical computation part 137.

In step S203, user's statistical computation part 137 is calculated the mean value of the evaluation mean value of the project that the user estimated.Specifically, user's statistical computation part 137 is selected the mean value paying close attention to the user and calculate the evaluation mean value of the project that this concern user estimated.For example, if user u1 pays close attention to the user, the project i2, the i3 that are estimated of user u1 and the evaluation mean value avg (Ri) of i5 are respectively 4.33,4.4 and 2.67 so.Therefore, the mean value of the evaluation mean value avg (Ri) of the project i2, the i3 that are estimated of user u1 and i5 is 3.8 (=(4.33+4.4+2.67)/3).The mean value of the evaluation mean value that user's statistical computation part 137 is calculated is set to pay close attention to user's reputation orientation index.User's statistical computation part 137 repeats this computing, becomes until all users and pays close attention to the user.

User's statistical computation part 137 will be used to represent that the information of each user's reputation orientation index offers information and presents part 142.Information presents part 142 the reputation orientation index that is obtained is added on each user's that user profile retaining part 144 kept the information.

In step S204, information presents part 142 the reputation orientation index is presented to the user.For example, when being used to show the order of the information relevant with user A by importation 121 input, information presents part 142 the reputation orientation index of user A is sent to display part 122 with other information.The reputation orientation index of display part 122 explicit user A and the information of the requesting users A of institute.

At this moment, the value of the reputation orientation index of explicit user A same as before perhaps can show the value that the mean value of the reputation orientation index by utilizing all users and the reputation orientation index regularization that variance makes user A obtain.In addition, if for example the reputation orientation index of user A surpasses predetermined threshold, show similar " you have high reputation orientation " such message so.

In this manner, the user who gives each project by effective use estimates, and the reputation orientation index that can obtain each user is for use in presenting.

In this, when described similar users is extracted the index of similarity that obtains in handling between the user with reference to Figure 13, use this reputation orientation index in the above.

Next, with reference to the process flow diagram among Figure 21, user characteristics (most index) computing that is used to calculate most indexes of representing class user statistics is described.

In step S221, user's cluster produces part 145 and produces user's cluster.At first, user's cluster produces part 145 and obtains the project evaluation history that historical retaining part 132 is kept.According to the project evaluation history of being obtained, user's cluster produce part 145 produce each users, for example its component is the matrix (be called user items hereinafter and estimate matrix) that gives the evaluation of estimate of each project.Estimate matrix by utilizing the user items that is produced, user's cluster generation part 145 is looked each user and is in the project space, and carries out user's cluster by utilizing within this project space such as the such preordering method of k-means (k-mean value) method.

Be used to make the data of user's cluster to be not limited to particular data.For example, also can use data such as other such type of user preference information.Employed here term " user preference information " is to be that the vector or the like of the metadata of the project estimated of user (for example having given the project of 4 minutes or 5 minutes according to 5 grades) is represented by its element.In this case, in this content metadata space, carry out user's cluster.

In addition, replace each user is categorized into user's cluster, for example, can use soft cluster method to obtain being used to represent that each user belongs to the ownership weighting of the ownership degree of each user's cluster.

Hereinafter, the situation that as shown in figure 22 5600 users is categorized into user's cluster 1 to 4 these 4 user's clusters is described.In the example of Figure 22,100 users belong to 1,4000 user of user's cluster and belong to 2,1000 users of user's cluster and belong to user's cluster 3, and 500 users belong to user's cluster 4.

User's cluster produces user that part 145 will be used to represent to belong to each each user's cluster, their user's cluster information of number or the like offers user's statistical computation part 137.

In step S222, user's statistical computation part 137 is calculated user's relative number.Specifically, user's statistical computation part 137 make belong to each each user's cluster number of users divided by the total number of users order to calculate the relative number of the user in each each user's cluster.For example, in the example of Figure 22, the relative number of the user in user's cluster 1 is 0.0179 (≌ 100/5600).

In step S223, the relative number of the user in the user's cluster under user's statistical computation part 137 each each user is set to most indexes of each each user.After this, user's statistical computation part 137 will be used to represent that the information of most indexes of each user offers information and presents part 142.Information presents part 142 the most indexes that obtained is added on each user's that user profile retaining part 144 kept the information.

In step S224, information presents part most indexes is presented to the user.For example, when having imported by importation 21 when being used to present the order of the information relevant with user A, information presents the information that part 142 will be used to represent most indexes of user A and is sent to display part 122 with the information of user A.The most indexes of display part 122 explicit user A and the information of the requesting users A of institute.

At this moment, for example, the value of most indexes of explicit user A same as before.Perhaps, if the value of most indexes of user A equals or is higher than predetermined threshold B, show similar " you be most " such message so,, show similar " you are minority " such message so if perhaps most indexes of user A are equal to or less than predetermined threshold C less than threshold value B.

In this manner, the user who gives each project by effective use estimates, and most indexes that can obtain each user are so that presented.

In this, when in described similar users extraction is handled with reference to Figure 13, obtaining the index of similarity between the user in the above, can use this majority index.

Next, with reference to the process flow diagram among Figure 23, user characteristics (prejudice index) computing that is used to calculate the prejudice index of representing class user statistics is described.

In step S241, project cluster generating unit divides 146 to produce the project cluster.Specifically, project cluster generating unit divides 146 to obtain the project evaluation history that historical retaining part 132 is kept.According to the project evaluation history of being obtained, project cluster generating unit divides 146 to produce the matrix (be called the project user hereinafter and estimate matrix) that its component of each project for example is the evaluation of estimate that gives of each user.By utilizing the project user who is produced to estimate matrix, project cluster generating unit divides 146 to look each project and be in the user's space, and comes the cluster of project implementation by utilizing within this user's space such as the such preordering method of k-means (k-mean value) method.Project cluster generating unit divides the project cluster information of 146 projects that will be used to represent to belong to each each project cluster, number of items or the like to offer user's statistical computation part 137.

Be used to make the data of project cluster to be not limited to particular data.For example, also can use the metadata of project.Under the situation of the metadata of utilizing project, be that the vector of metadata is represented each each project by its element, and in this metadata space the cluster of project implementation.

In addition, replacement as a project cluster, can use soft cluster method to obtain being used to represent that each project belongs to the ownership weighting of the ownership degree of each project cluster each classification of the items precedent.

Hereinafter, to becoming the situation of project cluster 1 to 4 these 4 project clusters to be described 1200 classifications of the items as shown in figure 24.In the example of Figure 24,200 projects belong to 1,450 project of project cluster and belong to 2,250 projects of project cluster and belong to project cluster 3, and 300 projects belong to project cluster 4.

In step S242, user's statistical computation part 137 is calculated the relative number of the evaluation that the user gave according to the project cluster.Specifically, at first, user's statistical computation part 137 is obtained the project evaluation history that historical retaining part 132 is kept.User's statistical computation part 137 selects one to pay close attention to the user, and according to the project evaluation history of being obtained, and will pay close attention to the number of items that the user estimates according to the project cluster and make form.After this, according to tabulating result, user's statistical computation part 137 is calculated the relative number of following evaluations, and this estimates the ratio that project that relative number represents that this concern user is estimated belongs to each each project cluster.

For example, consider to make the such situation of form according to each number of items that user u10 is estimated of 4 project clusters shown in Figure 24 as shown in figure 25.That is to say that in the project that user u10 is estimated, 15 projects belong to 1,40 project of project cluster and belong to 2,10 projects of project cluster and belong to project cluster 3, and 20 projects belong to project cluster 4.

At first, for each each project cluster, user's statistical computation part 137 obtains user u10 project of being estimated and the ratio that belongs to the project of each project cluster.For example, the project that user u10 is estimated is 0.075 (=15/200) with the ratio that belongs to the project of project cluster 1, the project that user u10 is estimated is 0.0889 (=40/450) with the ratio that belongs to the project of project cluster 2, the project that user u10 is estimated is 0.04 (=10/250) with the ratio that belongs to the project of project cluster 3, and user u10 project of being estimated and the ratio that belongs to the project of project cluster 4 are 0.0667 (=20/300).

Next, user's statistical computation part 137 is by carrying out regularization so that make the ratio summation of each project cluster that is obtained become 1, and obtains the relative number of evaluation for each project cluster.For example, the relative number of evaluation that has obtained the user u10 for project cluster 1 is 0.277 (≌ 0.075/ (0.075+0.0889+0.04+0.0667)).Similarly, the relative number of evaluation that has obtained for project cluster 2 is 0.329 (≌ 0.0889/ (0.075+0.0889+0.04+0.0667)), the relative number of evaluation that has obtained for project cluster 3 is 0.148 (≌ 0.04/ (0.075+0.0889+0.04+0.0667)), and the relative number of evaluation that has obtained for project cluster 4 is 0.246 (≌ 0.0667/ (0.075+0.0889+0.04+0.0667)).

That is to say that this estimates the ratio that project that relative number represents that user u10 is estimated belongs to each each project cluster, eliminated influence simultaneously the prejudice of the number of items that belongs to each project cluster.

Figure 26 has provided user u11 number of items of being estimated and the example of estimating the distribution of relative number purpose.For example, in Figure 26, in the project that user u11 is estimated, 90 projects belong to project cluster 1, and the relative number of the evaluation for project cluster 1 is 0.842.

In step S243, user's statistical computation part 137 is calculated the cluster prejudice (prejudice index) of the project that the user estimated.For example, user's statistical computation part 137 is calculated and is paid close attention to evaluation of user relative number purpose variances with as the prejudice index.For example, the evaluation relative number purpose variance of user u10 shown in Figure 10 is 0.00434, and promptly the prejudice index is 0.00434, and the evaluation relative number purpose variance of user u11 shown in Figure 11 is 0.117, and promptly the prejudice index is 0.117.

This prejudice exponential representation is to the prejudice degree of the project cluster specific distribution of the number of items that the user estimated.For example, if project is a video content, the so big described user of prejudice exponential representation is worthy of careful study to watching or listening to these projects with special characteristic very much.On the other hand, the little described user of prejudice exponential representation watches all items equably, and does not have very strong taste thus.

In addition, for example also can calculate the prejudice index by the function that utilizes monotone decreasing for estimating relative number purpose entropy.

The processing of user's statistical computation part 137 repeating step S242 and S243 becomes the concern user until all users, thereby calculates each user's prejudice index.After this, user's statistical computation part 137 will be used to represent that the information of each user's prejudice index offers information and presents part 142.Information presents part 142 the prejudice index that is obtained is added on each user's that user profile retaining part 144 kept the information.

In step S244, information presents part 142 the prejudice index is presented to the user.For example, when being used to present the order of the information relevant with user A by importation 121 input, information presents part 142 the prejudice index of user A is sent to display part 122 with other information.The prejudice index of display part 122 explicit user A and the information of the requesting users A of institute.

At this moment, for example, the value of the prejudice index of explicit user A same as before.Perhaps, if the prejudice index of user A equals or is higher than predetermined threshold B, can show similar " you are very special people " such message so, if perhaps the prejudice index of user A is lower than the predetermined threshold C less than threshold value B, can show similar " you have extensive hobby " such message so.

In this manner, the user who gives each project by effective use estimates, and the prejudice index that can obtain each each user is so that presented.

In this, when described similar users is extracted the index of similarity that obtains in handling between the user with reference to Figure 13, can use this prejudice index in the above.

Next, with reference to the process flow diagram among Figure 27, represent user characteristics (group the represents index) computing of index to be described to being used to calculate the group that represents class user statistics.

In step S261, identical with the processing of step S241 among above-mentioned Figure 23, project cluster generating unit divides 146 to produce the project clusters.Project cluster generating unit divides 146 will be used to represent that the project cluster information of the project cluster that produced offers user's statistical computation part 137.Hereinafter, to above-mentioned as shown in figure 24 become the situation of project cluster 1 to 4 these 4 project clusters to be described 1200 classifications of the items.

In step S262, user's statistical computation part 137 is made form according to the project cluster with the total number of all evaluation of user.Specifically, at first, user's statistical computation part 137 is obtained the project evaluation history that historical retaining part 132 is kept.According to the project evaluation history of being obtained, user's statistical computation part 137 is made form with all users to the total number of the evaluation of all items according to the project cluster.This tabulating result has provided the user to the interested indication of which project cluster.

Hereinafter, to the evaluation total number of as shown in figure 28 project cluster 1 be 1100, the evaluation total number of project cluster 2 is 5500, the evaluation total number of project cluster 3 be 2500 and the evaluation total number of project cluster 4 be that 2800 such situations are described.

In step S263, the index of similarity between the distribution of user's statistical computation part 137 calculating number of items that the user estimated and the distribution of all evaluation of user total numbers.For example, user's statistical computation part 137 selects one to pay close attention to the user, and will pay close attention to the number of items that the user estimates according to the project cluster and make form.After this, it is to be distance (for example cosine index of similarity, Euclidean distance or the like) between the vector of all evaluation of user total numbers of being broken according to the project cluster according to the break vector of the number of items that this concern user of (broke down) estimated and its element of project cluster that user's statistical computation part 137 is calculated its elements, with as number of items that the user was estimated in the distribution on the project cluster and all evaluation of user total numbers the index of similarity between the distribution on the project cluster.That is to say the index of similarity between the project cluster specific distribution of the project cluster specific distribution of user's statistical computation part 137 calculating concern number of items that the user estimated and the evaluation number of the whole group under the concern user.

For example, if the cosine index of similarity is used as index of similarity, the distribution (Figure 25) of the evaluation number of so above-mentioned user u10 and the distribution (Figure 28) of all evaluation of user total numbers are 0.976, and the distribution (Figure 28) of the distribution (Figure 26) of the evaluation number of above-mentioned user u11 and all evaluation of user total numbers is 0.291.

If this index of similarity is very high, this means the evaluation trend and concern evaluation of user similar trend of paying close attention to the affiliated whole group of user so.Therefore, pay close attention to the representative of consumer that the user is this group as can be known.On the contrary, if this index of similarity is very low, pays close attention to the user so as can be known and have the evaluation trend different with whole group.Therefore, as can be known user u10 more can representative of consumer u10 than user u11 and user u11 under group.Hereinafter, this index of similarity is called group and represents index.

When the project cluster specific distribution of the evaluation number that obtains whole group, not to use all evaluation of user total numbers.For example, can extract the user of predetermined number from this group randomly, and can use the evaluation of user total number that is extracted.

User's statistical computation part 137 double countings concern user's group represents the processing of index, become until all users and pay close attention to the user, thereby the group that calculates each user represents index.After this, user's statistical computation part 137 will be used to represent that on behalf of the information of index, each user's group offer information and present part 142.Information presents part 142 and represents index to add on each user's that user profile retaining part 144 kept the information group that is obtained.

In step S264, information presents part 142 and represents index to present to the user group.For example, when being used to present the order of the information relevant with user A by importation 121 input, information presents part 142 and represents index to be sent to display part 122 with out of Memory the group of user A.The group of display part 122 explicit user A represents the information of index and the requesting users A of institute.

At this moment, for example, the group of explicit user A represents the value of index same as before.Perhaps, if on behalf of index, the group of user A equal or is higher than predetermined threshold B, can show similar " you are the representative of consumer of this group " such message so, if perhaps on behalf of index, the group of user A be lower than predetermined threshold C less than threshold value B, can show similar " you are unusual in this group " such message so.

In this manner, the user who gives each project by effective use estimates, and the group that can obtain each each user represents index so that presented.

In this, when described similar users is extracted the index of similarity that obtains in handling between the user with reference to Figure 13, can use this group to represent index in the above.

Next, with reference to the process flow diagram among Figure 29, each represents that all user characteristics (index of conformity/current old index of the modish index/oneself) computing of the index of conformity of class user statistics, modish index and the current old index of oneself is described to being used to calculate it.

In step S281, identical with the processing of step S241 among above-mentioned Figure 23, project cluster generating unit divides 146 to produce the project clusters.At first, project cluster generating unit divides 146 will be used to represent that the project cluster information of the project cluster that produced offers user's statistical computation part 137.Hereinafter, to above-mentioned as shown in figure 24 be that the situation of project cluster 1 to 4 these 4 project clusters is described with 1200 classifications of the items.

In step S282, user's statistical computation part 137 is made form according to the project cluster and for each period with the number of items that the user estimated.Specifically, at first, user's statistical computation part 137 is obtained the project evaluation history that historical retaining part 132 is kept.User's statistical computation part 137 selects one to pay close attention to the user, and according to the project evaluation history of being obtained, according to the project cluster and will pay close attention to the number of items that the user estimates for each scheduled time slot and make form.

Employed term " period " is meant according to such as January, February or the March determined period of such absolute reference (being called the absolute period hereinafter) in this context, and the time of bringing into use service with the issuing time or the user of project is irrelevant.In addition, the length of this absolute period can be set to the common equal length of all users (for example month), perhaps length that can this absolute period for each each user is set to the period before the project of predetermined number is made evaluation.In the latter case, the period is different with the length of period.

Figure 30 to 32 has provided the distribution of the number of items of being estimated according to the user u20 to u22 that cluster broke in the absolute period 1 to 3.For example, in Figure 30, the number of items that belongs to project cluster 1 in the project that user u20 is estimated in the absolute period 1 is 5, the number of items that belongs to project cluster 1 in the project that user u20 is estimated in the absolute period 2 is 5, and the number of items that belongs to project cluster 1 in the project that user u20 is estimated in the absolute period 3 is 0.In addition, for example, in Figure 31, the number of items that belongs to project cluster 2 in the project that user u21 is estimated in the absolute period 1 is 40, the number of items that belongs to project cluster 2 in the project that user u21 is estimated in the absolute period 2 is 5, and the number of items that belongs to project cluster 2 in the project that user u21 is estimated in the absolute period 3 is 0.In addition, for example, in Figure 32, the number of items that belongs to project cluster 3 in the project that user u22 is estimated in the absolute period 1 is 30, the number of items that belongs to project cluster 3 in the project that user u22 is estimated in the absolute period 2 is 20, and the number of items that belongs to project cluster 3 in the project that user u22 is estimated in the absolute period 3 is 10.

In step S283, user's statistical computation part 137 is calculated the variability index of the distribution of assessment item number.That is to say that user's statistical computation part 137 is calculated the timing variations degree that is broken according to the project cluster, pay close attention to the distribution of the number of items that the user estimated.For example, utilization is represented as the distribution of assessment item number vector in the Item Sets group space, in each each period, and user's statistical computation part 137 is calculated cosine index of similarity between each vectors with the variability index as the distribution of assessment item number.

For example, under the situation of user u20, the cosine index of similarity between cosine index of similarity between the cosine index of similarity between absolute period 1 and absolute period 2, absolute period 2 and absolute period 3 and absolute period 1 and absolute period 3 is respectively 0.981,0.975 and 0.994.In addition, under the situation of user u21, the cosine index of similarity between cosine index of similarity between the cosine index of similarity between absolute period 1 and absolute period 2, absolute period 2 and absolute period 3 and absolute period 1 and absolute period 3 is respectively 0.288,0.638 and 0.0111.In addition, under the situation of user u22, the cosine index of similarity between cosine index of similarity between the cosine index of similarity between absolute period 1 and absolute period 2, absolute period 2 and absolute period 3 and absolute period 1 and absolute period 3 is respectively 0.464,0.359 and 0.0820.

The cosine index of similarity is high more, and the timing variations of the distribution that is broken according to the project cluster, pay close attention to the number of items that the user estimated is more little, and this pays close attention to the user and tends to come in the same manner project is made evaluation to consistance in each period.Hereinafter, this cosine index of similarity is called index of conformity.That is to say that index of conformity is represented the sequential stability index of distribution that broken according to the project cluster, the concern number of items that the user estimated.Similarity measurement except the cosine index of similarity also can be used as index of conformity.

In step S284, whether the distribution of user's statistical computation part 137 definite number of items of being estimated has changed.For example, be equal to or less than predetermined threshold (for example 0.5) if in all periods, pay close attention to user's index of conformity, user's statistical computation part 137 has been determined the changes in distribution of the number of items estimated so, otherwise determines that the distribution of the number of items estimated is stable.Perhaps, for example, equal or be higher than predetermined threshold (for example 0.9) if pay close attention to user's index of conformity in all periods, user's statistical computation part 137 determines that the distribution of the number of items estimated is stable so, otherwise has determined the changes in distribution of the number of items estimated.For example, under the situation of user u20 to u22, the distribution of definite number of items of estimating is stable for user u20, and for each of user u21 and u22 the changes in distribution of definite number of items of estimating.

For example, can be used for above-mentioned definite threshold value in advance and be set to appropriate value, this threshold value was changed for each period.

If user's statistical computation part 137 is determined to have paid close attention to the changes in distribution of the number of items that the user estimated, handle forwarding step S285 to so.

In step S285, user's statistical computation part 137 is made form according to the project cluster and for each period with all evaluation of user total numbers.Specifically, according to project evaluation history, user's statistical computation part 137 is made form according to the project cluster and for each scheduled time slot with all evaluation of user total numbers.For each period, this tabulating result has provided the user to the interested indication of which project cluster.

Figure 33 has provided the distribution of all evaluation of user total numbers that broken according to the project cluster in the absolute period 1 to 3.For example, in Figure 33, evaluation total number to the project that belongs to project cluster 1 in the absolute period 1 is 500, and the evaluation total number to the project that belongs to project cluster 1 in the absolute period 2 is 4000, and the evaluation total number to the project that belongs to project cluster 1 is 500 in the absolute period 3.

In step S286, for each period, the index of similarity between the distribution of user's statistical computation part 137 calculating number of items that the user estimated and the distribution of all evaluation of user total numbers.That is to say that user's statistical computation part 137 is calculated the group that pays close attention to the user for each period and represented index.

For example, if represent index by using the cosine index of similarity to calculate group, so in the absolute period 1 group of user u21 to represent index be 0.999, in the absolute period 2 group of user u21 represent index be 0.987, and in absolute period 3 group of user u21 to represent index be 1.000.On the other hand, in the absolute period 1 group of user u22 represent index be 0.269, in the absolute period 2 group of user u22 to represent index be 0.326, and in absolute period 3 group of user u22 to represent index be 0.325.

If this group represents exponential average very high, pay close attention to the user so as can be known and be have with world's (pay close attention to user under group) trend consistently change his/her user of trend trend of behavior (for example, watch or listen to which project).On the contrary, if this group represents exponential average very low, pay close attention to the user so as can be known and be the user with the current old type trend of oneself, it is indifferent to the behavior of the user except him, and he the type of interested project along with the time changes.

In this, make to have equaling or to be higher than 0.9 group that represent the user of the mean value of index be that new tidal stencils user and having is equal to or less than 0.4 group to represent the user of the mean value of index be self-current old type user if set in advance, so user u21 is categorized into new tidal stencils user, and user u22 is categorized into the current old type user of oneself.Hereinafter, represent the sequential mean value of index to be called modish index group, and inverse that will this modish index is called the current old index of oneself.

After this, this processing forwards step S287 to.

On the other hand, if determine that in step S284 the distribution of the concern number of items that the user estimated is stable, the processing of skips steps S285 and S286 so, and processing forwards step S287 to.

The processing of user's statistical computation part 137 repeating step S282 to S286 becomes the concern user until all users, thereby calculates each user's index of conformity and modish index.Yet, it should be noted that the processing of execution in step S285 at every turn is unless the user has changed the central tabulation period.After this, user's statistical computation part 137 will be used to represent that the information of each user's index of conformity and modish index offers information and presents part 142.Information presents part 142 index of conformity obtained and modish index is added on each user's that user profile retaining part 144 kept the information.

In step S287, information presents part 142 index of conformity and modish index is presented to the user.For example, when being used to present the order of the information relevant with user A by importation 121 input, information presents part 142 index of conformity and the modish index of user A is sent to display part 122 with other information.The information of the index of conformity of display part 122 explicit user A and modish index (perhaps self-current old index) and the requesting users A of institute.

At this moment, for example, the value of the index of conformity of explicit user A and modish index perhaps can show the feature (consistent type, new tidal stencils or the current old type of oneself) by index of conformity and the determined user A of modish index same as before.

In this manner, the user who gives each project by effective use estimates, and can obtain each user's index of conformity and modish index (perhaps self-current old index) so that presented.

In this, when described similar users is extracted the index of similarity that obtains in handling between the user with reference to Figure 13, can use these indexs of conformity and modish index in the above.

Next, with reference to the process flow diagram among Figure 34, to calculate its each all represent the instantaneous index of intermediate item statistics, the index that passes from mouth to mouth, designation number and fixedly the item characteristic of fan's index (the instantaneous index/index/designation number that passes from mouth to mouth/fixedly fan's index) computing be described.

In step S301, project statistical computation part 133 will be made form to the timing variations of the evaluation number of all items.Specifically, project statistical computation part 133 is obtained the project evaluation history that historical retaining part 132 is kept.According to this project evaluation history, project statistical computation part 133 will be made form for the evaluation number that each user of each period gives each project.

At employed term " period " in this context but be meant for the time point that becomes the time spent with respect to each each project, such as the so relative period (being called the relative period hereinafter) of first week, second week or the 3rd week that becomes in project after available.In addition, the length according to relative period of type of project is set to appropriate value.For example, if project is a music content, so because on the long a little period, sell music content, so the length of a period for example is set to one month.On the other hand, if project is the news article on the website, because the news article on the website has the instantaneity of height, therefore the length with a period for example is set to one day so.

In addition, project statistical computation part 133 will all users be made form to the evaluation total number of all items for each relative period.

Figure 35 has provided the example as a result of in the relative period 1 to 4 the evaluation number of project being tabulated.For example, in Figure 35, evaluation total number to all items in the relative period 1 is 53000, evaluation total number to all items in the relative period 2 is 30000, evaluation total number to all items in the relative period 3 is 4000, and the evaluation total number to all items is 3000 in the relative period 4.In addition, the evaluation number to project 1 in the relative period 1 is 500, and the evaluation number to project 1 in the relative period 2 is 100, and the evaluation number to project 1 in the relative period 3 is 15, and the evaluation number to project 1 is 10 in the relative period 4.

In step S302, the relative number of the evaluation at the previous period that is close to that project statistical computation part 133 was calculated in each each period.Specifically, for for each relative period of the second relative period, the ratio that project statistical computation part 133 is calculated the evaluation number in evaluation number and previous period that is being close in this relative period is with as the evaluation number for the previous period.

For example, Figure 36 has provided at the evaluation number for the previous period that tabulating result calculated among Figure 35.For example, in Figure 36, all items with regard to the relative period 1 in the relative period 2 for the previous period, be 0.57 (=30000/53000) for estimating number (being called the evaluation number in the relative period 2 hereinafter simply) with respect to the previous period, with regard to the relative period 2 in the relative period 3 is 0.13 (=4000/30000) with respect to the evaluation number for the previous period (being called the evaluation number for the previous period in the relative period 3 hereinafter simply), and in the relative period 4 is 0.75 (=3000/4000) with respect to the evaluation number for the previous period (being called the evaluation number for the previous period in the relative period 4 hereinafter simply) with regard to the relative period 3.In addition, in period 2, project 1 the relatively evaluation number for the previous period is 0.2 (=100/500), the evaluation number for the previous period in period 3 is 0.15 (=15/100) relatively, and the evaluation number for the previous period in the relative period 4 is 0.67 (=10/15).

In step S303, project statistical computation part 133 is calculated instantaneous index, the index that passes from mouth to mouth, designation number and fixing fan's index.Specifically, for example, as shown in figure 37, but can think when project become the time spent the most frequent evaluation and the project that its evaluation number shortly after that reduces had very high instantaneous index.For example, if project is a video content, appear on the market so suddenly, be project by the project of watching the most continually/listening to and shortly after that stopping to watch/listen at first with high instantaneous index.

Project statistical computation part 133 is according to the instantaneous index of the evaluation number of this project being determined each each project with respect to the degree of the quick reduction of average tendency of all items.For example, in the example of Figure 36, relatively the mean value of the evaluation number for the previous period of all items in period 2 and relative period 3 is 0.35, and the mean value of the evaluation number for the previous period of the project 1 in relative period 2 and relative period 3 is 0.18.Therefore, as can be known the evaluation number of project 1 speed with the average velocity twice of about all items is reduced.

In this case, the instantaneous index of acquisition project 1 be 1.9 (=0.35/0.18), its be by make project 1 in relative period 2 and relative period 3 with respect to the mean value of the evaluation number for the previous period the value that mean value obtained divided by the evaluation number for the previous period of all items in period 2 relatively and relative period 3.That is to say, but instantaneous exponential representation is estimated for the average velocity that number reduces the relative velocity that the evaluation number to each each project reduces with respect to become the time spent from project.

As shown in figure 38, the initial project of still progressively frequently not estimated by a large amount of evaluations is an intermediate item of word of mouth.This as can be known project has the height index that passes from mouth to mouth.For example, be under the situation of video content in project, if the sale number of watching/listen to the number of times of this project or project is at leisure but stably increase, this project has the height index that passes from mouth to mouth so as can be known.For example, in the example of Figure 36, the evaluation number for the previous period of project 2 is 1 or bigger in all relative periods 2 to 4, and very big and have a value of 3.3 in relative period 3.Therefore, suppose that project 2 progressively but stably increases after its issue, and after this in the relative period 3, increasing rapidly.

For example, be set to the index that passes from mouth to mouth by the value that makes from the relative period 2 to relative periods all evaluation numbers for the previous period in 4 multiply each other together to be obtained.In this case, the index that passes from mouth to mouth of project 2 is 5.35 (=1.2 * 3.3 * 1.35).Perhaps, for example, only under the situation of evaluation number in the last period relatively of tabulation within the period greater than the evaluation number in the first relative period, the index that can pass from mouth to mouth be set to by make from wherein be last time with respect to the evaluation number for the previous period 1 or littler relative period after the relative period extremely wherein be last time with respect to the evaluation number for the previous period 1 or bigger relative period in, value that evaluation number for the previous period multiplies each other together to be obtained.Therefore, the pass from mouth to mouth length of the period of exponential representation during the evaluation number of each each project is increased and the increase degree of estimating number.

In addition, for example, as shown in figure 39, as can be known according to stationary mode estimate and with irrelevant project of time be project with high standard index.For example, be under the situation of video content in project, if watching/listening to or selling this project according to stationary mode in the long duration very much, this project has high designation number so.That is to say, as can be known when the mean value m of evaluation number for the previous period becomes near 1, it changes σ ²Diminish and satisfy these conditions period, p became the time, designation number uprises.Therefore, for example, can be by p * N (m; 1, σ ²) define designation number.Function N () is the probability density function by the represented normal distribution of following equation (8).

$N(x;\mathrm{\&mu;},{\mathrm{\&sigma;}}^2)=\frac{1}{\sqrt{2\mathrm{\&pi;\&sigma;}}}\exp (-\frac{{(x-\mathrm{\&mu;})}^2}{{2\mathrm{\&sigma;}}^2})\&CenterDot;\&CenterDot;\&CenterDot;\left(8\right)$

Period p is set to the following period, evaluation number during this period for the previous period (for example drops on preset range constantly, 0.8 to 1.2) within, and the evaluation during this period in each corresponding period has relatively outnumbered following predetermined threshold, this predetermined threshold or on the time think that project is a standards project.

Under the situation of the project 3 in Figure 36, the mean value m of the evaluation number for the previous period in the period 2 to 4 equals 0.98 relatively, and it changes σ ²Equal 0.012, so designation number is 10.7 (=3 * { 1/ (2 π * 0.012) ^0.5* exp ((0.98-1) ²/ (2 * 0.012)) }).Therefore, designation number is represented the sequential stability index to the evaluation number of each each project.

In addition, suppose in the project with high standard index, by the user of close limit especially the project often estimated be to have fixedly fan's project.

Figure 40 has provided the transformation to the evaluation number of project 3 in the relative period 1 to 4, and Figure 41 has provided the transformation to the evaluation number of project 4 in the relative period 1 to 4.Evaluation total number in each relative period is identical for project 3 with project 4.Yet, it should be noted that 100 users altogether from user 1001 to 1100 in the relative period 1 to 4 make evaluation to project 3, and from 20 users altogether of user 2001 to 2020 project 4 are made evaluation.In this case, be the average ratings number of each user within scheduled time slot with fixing fan's index definition.Therefore, the fixedly fan index of the project 3 in the period 1 is 1.2 (=120/100) relatively, and the fixedly fan index of project 4 is 6 (=120/20).For example, be under the situation of video content in project, if specific people watches/listens to this project for a long time or sells specific people with this project for a long time, this project has high fixedly fan index so.

Project statistical computation part 133 repeats to select a concern project and obtains the instantaneous index of this concern project, the index that passes from mouth to mouth, designation number and the fixedly such processing of fan's index, become the concern project until all items, thereby obtain the instantaneous index of each project, the index that passes from mouth to mouth, designation number and fixing fan's index.Project statistical computation part 133 will be used to represent the instantaneous index of each project of being obtained, the index that passes from mouth to mouth, designation number and fixedly the information of fan's index offer information and present part 142.Information present part with the instantaneous index of each project of being obtained, the index that passes from mouth to mouth, designation number and fixedly fan's index add on the information of each project that project information retaining part 143 kept.

At this moment, can with the instantaneous index of each project of being obtained, the index that passes from mouth to mouth, designation number and fixedly fan's index offer item types determining section 134 to determine the item types of each project from project statistical computation part 133.For example, with its instantaneous index, the index that passes from mouth to mouth, designation number and fixedly fan's index item types of surpassing the project of corresponding predetermined threshold be defined as instantaneous type respectively, the type that passes from mouth to mouth, standard form and fixing fan's type.

In step S304, information present part 142 with instantaneous index, the index that passes from mouth to mouth, designation number and fixedly fan's index present to the user.For example, when as in the processing procedure at the step S1 of Fig. 4, the information relevant with project being presented to the user, information present the instantaneous index that part 142 also will be used to represent this project, the index that passes from mouth to mouth, designation number and fixedly the information of fan's index be sent to display part 122.Display part 122 shows the instantaneous index of these projects, the index that passes from mouth to mouth, designation number and fixing fan's index and the information relevant with this project that this user asked.

At this moment, can show the instantaneous index of this project, the index that passes from mouth to mouth, designation number and the fixing value of fan's index same as before, perhaps can show according to instantaneous index, the index that passes from mouth to mouth, designation number and the fixedly indication of the determined item types of fan's index, that is to say, can show instantaneous type, the type that passes from mouth to mouth, standard form and fixing fan's type.

In this manner, the user who gives each project by effective use estimates, can obtain the instantaneous index of each project, the index that passes from mouth to mouth, designation number and fixedly fan's index so that present to the user.Therefore, the user can learn the evaluation trend that gives each project exactly.

Next, with reference to the process flow diagram among Figure 42, to be used to calculate its each all represent the faddict B index, connoisseur's index, conservative index of class user statistics and fixedly the user characteristics of fan's index (faddict B index/connoisseur's index/conservative index/fixedly fan's index) computing be described.

In step S321, user's statistical computation part 137 is obtained the feature of the project that the user estimates.Specifically, user's statistical computation part 137 selects one to pay close attention to the user, and obtains the project evaluation history subscriber-related with this concern from historical retaining part 132.In addition, user's statistical computation part 137 presents part 142 is obtained the feature (instantaneous index, the index that passes from mouth to mouth, designation number and fixedly fan's index) of the project that is used to represent that this concern user is estimated from project information retaining part 143 information by information.

In step S322, user's statistical computation part 137 is calculated users' faddict B index, connoisseur's index, conservative index and fixing fan's index.For example, estimate project continually, so therefore can new feature that pay close attention to the user be defined with special characteristic if pay close attention to the user.In this case, according to the ratio of the project with special characteristic and the total number of paying close attention to the project that the user estimated, to the evaluation total number and ratio of paying close attention to the evaluation of user total number or the like of project, determine whether to have estimated continually project with special characteristic with special characteristic.In this case, being defined as and paying close attention to the user and repeatedly identical items is made evaluation if will estimate total number, whenever this project being made when estimating, is an evaluation with its counting so.

For example, in general, usually strengthen identification in advance to project with high instantaneous index by advertisement or the like.Therefore, pay close attention to the user after available project with high instantaneous index is estimated if be right after to become in project, paying close attention to the user so as can be known is the faddict.In the following description, for above distinguishing with reference to faddict's index of faddict's index of the described project-based main index of Figure 10 or the like and project-based instantaneous index as described below, the former is called faddict A index, and the latter is called faddict B index.

For example, suppose that 40 projects paying close attention in the project that the user estimated are the instantaneous type projects with the instantaneous index that is equal to or higher than predetermined threshold, if within the period 1 80 projects are being made evaluation relatively, so 0.4 * 0.8=0.32 are being defined as the faddict B index of paying close attention to the user.That is to say that faddict B index is based on the ratio of the instantaneous type project of being estimated within the scheduled time slot that becomes in project after available with the project that the user estimated of concern.At this moment, estimate faddict B index period needn't with employed relative period unanimity when the instantaneous index of project is estimated.For example, can in the period of shorter more segmentation, estimate faddict B index.In addition, for example, in order to reduce the influence of instantaneous type project and the ratio of paying close attention to the project that the user estimated, with (0.4) ^0.5* 0.8=0.51 is defined as faddict B index.

Hence one can see that, and on the contrary, the user who has low faddict B index after a period of time passes estimates instantaneous type project, and thereby as can be known this user chase popular popular follower (hit follower) type user.

In addition, for example, pay close attention to the user after available and estimate having the pass from mouth to mouth project of index of height if be right after to become in project, paying close attention to the user so as can be known is the connoisseur user that trend is predicted.

For example, suppose that 40 projects paying close attention in the project that the user estimated are the type that the passes from mouth to mouth projects that have equaling or be higher than the index that passes from mouth to mouth of predetermined threshold, if within the period 1 80 projects are being made evaluation relatively, 0.4 * 0.8=0.32 can be defined as connoisseur's index of paying close attention to the user so.That is to say that this connoisseur's index is based on the ratio of the type that the passes from mouth to mouth project of being estimated within the scheduled time slot that becomes in project after available with the project that the user estimated of concern.At this moment, pay close attention to time that the user makes evaluation to this given project more early with respect to wherein the evaluation total number of given project being become the highest relative period as can be known, connoisseur's rank of concern user is high more.In addition, the period that connoisseur's index is estimated needn't with employed relative period unanimity when the index that passes from mouth to mouth of project is estimated.For example, can be at the shorter more period inner evaluation connoisseur index of segmentation.In addition, for example, has the pass from mouth to mouth influence of project with the ratio of paying close attention to the project that the user estimated of index of height in order to reduce, with (0.4) ^0.5* 0.8=0.51 is defined as connoisseur's index.

Hence one can see that, and on the contrary, the user who has low connoisseur's index after a period of time passes estimates the type project of passing from mouth to mouth, and thereby as can be known this user chase popular follower's type user of passing from mouth to mouth.

In addition, for example, only mainly the project with high standard index is estimated, paid close attention to the user so as can be known and guard if pay close attention to the user.For example, can be same as before the ratio of standard form project and the project that the user estimated of concern that has equaling or be higher than the designation number of predetermined threshold be defined as conservative index.

In addition, for example, only mainly the project with higher fixedly fan index is estimated if pay close attention to the user, paying close attention to the user so as can be known is the fixedly fan of specific project.For example, can be same as before the number ratio of fixedly fan type project and the project that the user estimated of concern that has equaling or be higher than the fixedly fan index of predetermined threshold be defined as fixedly fan's index.

User's statistical computation part 137 repeats to select the faddict B index paying close attention to the user and obtain this concerns user, connoisseur's index, guards index and the fixedly processing of fan's index, become until all users and to pay close attention to the user, thereby obtain each user's faddict B index, connoisseur's index, conservative index and fixing fan's index.User's statistical computation part 137 will be used to represent each user's of being obtained faddict B index, connoisseur's index, conservative index and fixedly the information of fan's index offer information and present part 142.Information present part with each user's of being obtained faddict B index, connoisseur's index, conservative index and fixedly fan's index add on each user's that user profile retaining part 144 kept the information.

In step S323, information present part 142 with faddict B index, connoisseur's index, conservative index and fixedly fan's index present to the user.For example, when being used to present the order of the information relevant with user A by importation 121 input, information present part 142 with the faddict B index of user A, connoisseur's index, conservative index and fixedly fan's index add on the information of user A and and be sent to display part 122 this information.The faddict B index of display part 122 explicit user A, connoisseur's index, conservative exponential sum be fan's index and the information relevant with user A that the user asked fixedly.

In this manner, according in the middle of the item characteristic of project statistics (instantaneous index, the index that passes from mouth to mouth, designation number and fixedly fan's index) representative, pay close attention to the feature that numerous items had that the user estimated, user's statistics (faddict B index, connoisseur's index, conservative index and fixedly fan's index) that can obtain to pay close attention to the user is so that present to the user.

Now, with reference to Figure 43 and Figure 44, to described user characteristics and item characteristic gather hereinbefore.

Figure 43 is the form that is used to gather item characteristic.According to the raw data that is used to obtain item characteristic, item characteristic roughly is divided into three groups.

As top described, represent the feature that obtained according to project evaluation history for first group with reference to figure 4 or the like.This group comprises main index, estimates mean value and estimates variance.

As top described, represent for second group to add up the feature that is obtained according to the project that includes main index, evaluation mean value and evaluation variance with reference to figure 4,17 or the like.This group comprises masterpiece, hiding masterpiece, dispute works, enthusiast's appealing, useless works, merits no attention, produces in a large number works and rough works.

Represent according to the feature that timing transition obtained of estimating number for the 3rd group.This group comprises instantaneous type, the type that passes from mouth to mouth, standard form and fixing fan's type.

Because the summary of each each feature is described above, therefore the descriptions thereof are omitted to avoid repetition.

Figure 44 is used for form that user characteristics is gathered.Roughly user characteristics is divided into four groups, comprising the feature relevant with user's social status, with to the relevant feature of the trend of the user of contents of a project orientation, with relevant feature and the further feature of user's feeler that is used to catch fresh information.

The feature group relevant with user's social status comprises that faddict A index enthusiast's index of its opposite face (perhaps as), main orientation index pettifogger's index of its opposite face (perhaps as), most index the minority index of its opposite face (perhaps as), group represent index and modish the index self-current old index of its opposite face (perhaps as).

User with high faddict A index is such user: the evaluation number of main project with high main index is tended to very big, that is, tend to give the very user of the evaluation of big figure to main project.On the other hand, the user's (low faddict A index) with high enthusiast's index tends to give the very user of the evaluation of big figure to the less important project with low main index, promptly tends to give the very user of the evaluation of big figure to less important project.Therefore, the main correlation of indices of faddict A exponential sum enthusiast's index and project.

User with high main orientation index tends to the user that speaks highly of to main project.On the other hand, the user with high pettifogger's index (low main orientation index) tends to the user that speaks highly of to less important project.Therefore, the main correlation of indices of main orientation index and pettifogger's index and project.

User with high most indexes is the user who tends to belong to the user's cluster with a large number of users.On the other hand, the user with high minority index (low most indexes) is the user who tends to belong to the user's cluster with small number of users.

Having high group, to represent the user of index be such user: the distribution of the evaluation number that is broken according to the project cluster is tended to similar to all users' distribution.

User with high modish index is such user: the timing transition according to the distribution of the evaluation number of project cluster tends to synchronously change with all users' distribution.On the contrary, the user with current old index of high oneself (low modish index) is such user: the timing transition according to the distribution of the evaluation number of project cluster tends to synchronously change hardly with all users' distribution.

With the relevant feature group of the trend of the user of contents of a project orientation is comprised general index number and reputation orientation.

User with high general index number is such user: the evaluation of estimate of each each project is tended to have and the high correlation of estimating mean value.Therefore, general index number is relevant with the evaluation mean value of project.

User with high reputation orientation index tends to the user that estimates to the project with high praise mean value.Therefore, the reputation orientation index is relevant with the evaluation mean value of project.

The feature group relevant with the user's feeler that is used to catch fresh information comprises faddict B index popular follower's index of its opposite face (perhaps as) and connoisseur's index index that passes from mouth to mouth of its opposite face (perhaps as).

User with high faddict B index is the user who tends to from just estimating to the instantaneous type project with high instantaneous index in early days.On the other hand, the user with high popular follower's index (perhaps low faddict B index) is the user that can not tend to from just estimating to instantaneous type project in early days.Therefore, the instantaneous correlation of indices of popular follower's index of faddict B exponential sum and project.

User with high connoisseur's index be tend to project arouse attention and increase sharply popular before to having the height user that the type that the passes from mouth to mouth project of index estimates of passing from mouth to mouth.On the other hand, have height pass from mouth to mouth the user of follower's index (low connoisseur's index) be can not tend to project arouse attention and increase sharply popular before to having the height user that the type that the passes from mouth to mouth project of index estimates of passing from mouth to mouth.Therefore, connoisseur's exponential sum correlation of indices that passes from mouth to mouth of follower's index and project that passes from mouth to mouth.

Other feature groups comprise prejudice index, index of conformity and fixing fan's index.

User with high prejudice index is such user: the project that the user estimated is with prejudice to the specific project cluster consumingly.

User with high index of conformity is such user: the timing variations according to the distribution of the assessment item number of project cluster is tended to very little, that is, the user with high index of conformity is such user: can great changes will take place in time according to the distribution of the assessment item number of project cluster.

User with high conservative index is such user: the evaluation number of standard form project with high standard index is tended to very big, that is, tend to give the user that big figure is estimated very much to the standard form project.Therefore, conservative index is relevant with the designation number of project.

Having the high fixedly user of fan's index is such user: tend to very greatly to having the high fixedly evaluation number of the fixedly fan type project of fan's index, that is, tend to give the very user of big figure evaluation to fixing fan's type project.Therefore, the fixing fixedly fan correlation of indices of fan's index and project.

Next, with reference to Figure 45 to 54, the processing that information handling system 100 is presented to the user with the information relevant with project is described.

At first, with reference to the process flow diagram among Figure 45, the message block personalisation process is described.Message block is information to be presented to user's unit.In the following description, will in this processing, the user to its presentation information be called the concern user.

In step S401, information presents part 142 and obtains and present the rule that presents that regular retaining part 147 kept.This has presented rule definition branch condition from the processing that step S402 begins and the rule that is used for the display message piece.System supplier can freely change and presents rule.

In step S402, information presents part 142 and determines to pay close attention to the feature whether user has group 1.Specifically, information presents part 142 and obtains and pay close attention to subscriber-related information from user profile retaining part 144.If satisfy one of following condition: the faddict A index of paying close attention to the user equals or is higher than predetermined threshold; The faddict B index of paying close attention to the user equals or is higher than predetermined threshold; The main orientation index of paying close attention to the user equals or is higher than predetermined threshold; The modish index of paying close attention to the user equals or is higher than predetermined threshold; And the prejudice index of paying close attention to the user equals or is higher than predetermined threshold, and information presents part 142 and determines to pay close attention to the feature that users have group 1 so.This aftertreatment forwards step S403 to.

In step S403, information presents part 142 and presents advertisement.Specifically, information presents part 142 and produces the information relevant with advertisement for paying close attention to the user, and this information is sent to display part 122.Display part 122 is according to the information display ads that is obtained.After this, processing forwards step S404 to.

On the other hand, if in step S402, determine to pay close attention to the feature that the user does not have group 1, the processing of skips steps S403 so, and processing forwards step S404 to.

In step S404, information presents part 142 and determines to pay close attention to the feature whether user has group 2.Specifically, if one of meet the following conditions: the faddict A index of paying close attention to the user equals or is higher than predetermined threshold; The faddict B index of paying close attention to the user equals or is higher than predetermined threshold; The main orientation index of paying close attention to the user equals or is higher than predetermined threshold; Most indexes of paying close attention to the user equal or are higher than predetermined threshold; The modish index of paying close attention to the user equals or is higher than predetermined threshold; Popular follower's index of paying close attention to the user equals or is higher than predetermined threshold; And the follower's index that passes from mouth to mouth of paying close attention to the user equals or is higher than predetermined threshold, and information presents part 142 and determines to pay close attention to the feature that users have group 2 so.This aftertreatment forwards step S405 to.

In step S405, information presents part 142 and presents rank.Specifically, information presents part 142 and produces the information relevant with rank according to the evaluation number to each project, and this information is sent to display part 122.Display part 122 is according to the rank of the information display items display of being obtained.After this, processing forwards step S406 to.

On the other hand, if in step S404, determine to pay close attention to the feature that the user does not have group 2, the processing of skips steps S405 so, and processing forwards step S406 to.

In step S406, information presents part 142 and determines to pay close attention to the feature whether user has group 3.Specifically, if one of meet the following conditions: the faddict A index of paying close attention to the user is less than predetermined threshold; The modish index of paying close attention to the user is less than predetermined threshold (the current old index of oneself equal or be higher than predetermined threshold); And the prejudice index of paying close attention to the user is less than predetermined threshold, and information presents part 142 and determines to pay close attention to the feature that users have group 3 so.This aftertreatment forwards step S407 to.

In step S407, information presents part 142 recommendation list is presented to the concern user.Specifically, information presents part 142 and produces the tabulation that the recommended project of being extracted is handled in for example project recommendation from above-mentioned Figure 15 or 16 for paying close attention to the user, and this tabulation is sent to display part 122.According to the tabulation of being obtained, display part 122 shows recommendation list to paying close attention to the user.After this, processing forwards step S408 to.

On the other hand, if in step S406, determine to pay close attention to the feature that the user does not have group 3, the processing of skips steps S407 so, and processing forwards step S408 to.

In step S408, information presents part 142 and determines to pay close attention to the feature whether user has group 4.Specifically, if concern user's reputation orientation index equals or is higher than predetermined threshold, information presents the feature that part 142 definite users of concern have group 4 so.After this, processing forwards step S409 to.

In step S409, information presents part 142 and presents project evaluation information.Specifically, when title that presents given project and details, information presents part 142 statistics (for example estimating mean value) that gives the evaluation of this project is sent to display part 122 together with the information relevant with this project.Display part 122 also shows the evaluation statistics of being obtained when the title of the project of obtaining when presenting and details.After this, processing forwards step S410 to.

On the other hand, if in step S408, determine to pay close attention to the feature that the user does not have group 4, the processing of skips steps S409 so, and processing forwards step S410 to.

In step S410, information presents part 142 and determines to pay close attention to the feature whether user has group 5.Specifically, if connoisseur's index of concern user equals or is higher than predetermined threshold, information presents the feature that part 142 definite users of concern have group 5 so.After this, processing forwards step S411 to.

In step S411, information presents part 142 and presents the newcomer.Specifically, information presents part 142 and produces and the relevant information of project of setting up explicit evaluation not yet with regard to it, and this information is sent to display part 122.Display part 122 shows the information of being obtained as the information relevant with the newcomer.For example, under the situation of music distribution business, show and the relevant information of new artist of yet it not being set up explicit evaluation.After this, the message block personalisation process finishes.

On the other hand, if determine to pay close attention to the feature that the user does not have group 5 in step S410, the personalisation process of the processing of skips steps S411, and message block so finishes.

In this manner, can come selection information so that presented according to the feature that the user adds up represented concern user.

Except the message block of selecting to show according to the feature of paying close attention to the user as mentioned above, for example the display priority of message block, size or the like also can change.

Figure 46 provided according to above-mentioned message block personalisation process will be in the music distribution business to the example of the shown screen of the user with the high reputation orientation index of high faddict A exponential sum.By the processing among Figure 45, determine to have the feature that the user of the high reputation orientation index of high faddict A exponential sum has group 1, group 2 and organizes 4.Therefore, on the screen in Figure 46, show rank window 201 and the advertisement windows 202 that is used for the display items display rank to messagewindow 203 with the new of music content.

In addition, Figure 47 provided according to above-mentioned message block personalisation process will be in the music distribution business to the example of the shown screen of the user with the high connoisseur's index of the high current old exponential sum of oneself.By the processing among Figure 45, determine that the user with the high connoisseur's index of the high current old exponential sum of oneself has group 3 and organizes 5 feature.Therefore, on the screen in Figure 47, be used to show the recommendation list window 211 that the recommended project is tabulated and be used to show and newcomer's window 212 of the relevant information of popular newcomer of not increasing sharply as yet with new the demonstration together of music content to messagewindow 213.

Next, the process flow diagram with reference among Figure 48 is described filtration treatment.In the following description, will in this processing, the user to its presentation information be called the concern user.

In step S431, the recommended project is extracted part 141 and is created basic list.The recommended project is extracted the project that part 141 is complementary by query search or the like extraction and predetermined condition, and creates the tabulation of the project of extracting, and promptly creates basic list.For example, if project is a music content, the list of artists of music that create to play predetermined style (for example popular, jazz, classic or the like) so is with as basic list.

In step S432, the recommended project is extracted part 141 and select a project from basic list.Hereinafter, item selected is called the concern project thus.

In step S433, the recommended project is extracted part 141 and is determined whether this project has the feature that is complementary with the user.Specifically, recommended project extraction part 141 presents part 142 is obtained the concern user from user profile retaining part 144 user profile by information.The recommended project is extracted part 141 and is extracted according to the form among Figure 44 and the concern relevant item characteristic of feature that the user had.

In addition, recommended project extraction part 141 presents part 142 is obtained the concern project from project information retaining part 143 project information by information.According to the project information of being obtained, the recommended project extract part 141 obtain with the concern project in, the rank of each each item characteristic that concern each each feature that the user had is relevant.If the rank of the item characteristic that is obtained equals or is higher than predetermined threshold, the recommended project is extracted part 141 feature that definite concern projects have and the concern user is complementary so, and this aftertreatment forwards step S434 to.For example, paying close attention under the situation of feature that the user has faddict A (if faddict's index of user equals or is higher than predetermined threshold),, satisfy above-mentioned condition so if the main index of the project of concern equals or is higher than predetermined threshold.

In step S434, the recommended project is extracted part 141 the concern project is added in the new tabulation.After this, processing forwards step S435 to.

On the other hand, if the rank of each each item characteristic that in step S433, is obtained less than predetermined threshold, the recommended project is extracted part 141 and is determined that the concern projects are not to have and the project of paying close attention to the feature that the user is complementary so.After this, the processing of skips steps S434, and processing forwards step S435 to.

In step S435, the recommended project is extracted part 141 and is determined whether basic list is finished.If still exist in basic list not as the processed project of concern project, recommended project extraction part 141 definite basic lists are not finished so, and handle and get back to step S432.After this, the processing of repeating step S432 to S435 is finished until definite basic list, and extracts to have with the project of paying close attention to the feature that the user is complementary and with it from basic list and add in the new tabulation.

On the other hand, finish, handle forwarding step S436 to so if in step S435, determine basic list.

In step S436, information presents part 142 and will newly tabulate and present to the user.Specifically, the recommended project is extracted part 141 and the new tabulation that is produced is offered information is presented part 142.Information present part 142 from project information retaining part 143 obtain be included in new tabulation within the relevant information of project, and the information of being obtained is sent to display part 122.According to the information of being obtained, display part 122 show be included in new tabulation within the relevant information of project.After this, filtration treatment finishes.

For example, consider to pay close attention to the very high and basic list of user's faddict A exponential sum reputation orientation index and comprise the such situation of project 1 to 5 with feature as shown in figure 49.In the figure, the every tabulation with circle shows that the rank of respective item feature is very high.For example, project 1 has high main index, low mean value and the height index that passes from mouth to mouth of estimating.

In this case, according to the form among Figure 44, extract main index with as with the item characteristic of faddict A correlation of indices, and extract and estimate mean value with as the item characteristic relevant with the reputation orientation index.Therefore, extraction has the project 1,2,4 and 5 of high main index or high praise mean value in the basic list from Figure 49, and it is presented to the concern user as new tabulation.

In this manner, can extract have by project statistics represented and add up the project of the relevant feature of represented concern user's feature so that present to the user with the user.

If in new tabulation, even do not comprising any single project as carrying out the result that this item extraction handles, also can present be included in basic list within the relevant information of all items.

Next, with reference to the process flow diagram among Figure 50, the highlighted display process of item characteristic is described.In the following description, will in this processing, the user to its presentation information be called the concern user, and the project that presents the information relevant with it will be called the concern project.

In step S451, information presents part 142 and obtains project information.That is to say that information presents part 142 is obtained the concern project from project information retaining part 143 project information.Information presents part 142 project information of being obtained is sent to display part 122.

In step S452, the processing of extracting part 141 with the recommended project among the step S433 of above-mentioned Figure 48 is identical, and information presents part 142 and determines whether the concern projects have and pay close attention to the feature that the user is complementary.Have the feature that is complementary with the concern user if determine the concern project, handle forwarding step S453 to so.

In step S453, information presents the feature that part 142 indicated number parts, 122 highlighted demonstrations and user are complementary.Specifically, information presents part 142 will represent that the information of following item characteristic is sent to display part 122, and indicated number part 122 these item characteristics of highlighted demonstration, wherein said item characteristic are definite item characteristics that is complementary with the concern user that had by the concern project in step S452.

If determine that in step S452 the concern project does not have the feature that is complementary with the concern user, the processing of skips steps S453 so, and processing forwards step S454 to.

In step S454, the user is presented to project information in display part 122.That is to say that display part 122 shows the information relevant with the concern project.

Figure 51 provided according to the highlighted display process of above-mentioned item characteristic in the music distribution business to the example of the shown screen of the user with the high reputation orientation index of high faddict A exponential sum.In zone 221, show collection of records big envelope as the music content of the project of concern.In zone 222, the year, month, day and the item characteristic of the album title of demonstration concern project, artist name, style, distribution.In zone 223, show the commentary text of concern project.It is to have high main exponential sum height the pass from mouth to mouth main type of index and the type project of passing from mouth to mouth that project is paid close attention in the demonstration of zone in 222.

Now, according to the form among Figure 44, the item characteristic relevant with faddict A exponential sum reputation orientation index is that main exponential sum is estimated mean value.Therefore, in 222 in the shown item characteristic, come highlighted demonstration speech " mainly " in the zone with thick letter.This can make the concern user notice the concern project more.

In this manner, can be highlighted the display items display statistics represented and add up the relevant item characteristic of represented concern user's feature so that presented with the user.

If on the website, show the screen of Figure 51, so for example, can realize highlighted demonstration by category attribute being added on the label that includes item characteristic " mainly " and using style sheet (style sheet).

Next, with reference to the process flow diagram among Figure 52, popular prediction processing is described.In the following description, will be called the concern project to its project of carrying out this processing.

In step S471, project statistical computation part 133 is obtained the user's who estimates to project feature.For example, project statistical computation part 133 is obtained the project evaluation history relevant with the concern project from historical retaining part 132.According to the project evaluation history of being obtained, project statistical computation part 133 is extracted the user who has estimated to the concern project.At this moment, replace extracting all users that estimated, issued the user that the concern project was estimated within certain period afterwards but for example also can extract the user of predetermined number or be extracted in to the concern project.Project statistical computation part 133 presents part 142 by information and extracts the user's who is extracted user profile from user profile retaining part 144.Project statistical computation part 133 will have the extraction user's of each user characteristics ratio (being called owning rate hereinafter) and make form.

Among Figure 53 and 54 each has provided the example of owning rate of user characteristics of project 1 and project 2 being made the user of evaluation.For example, Figure 53 shows: project 1 is being made among the user of evaluation, have that its faddict A index equals or user's the ratio that is higher than the faddict A feature of predetermined threshold is 0.3, have that its faddict B index equals or user's the ratio that is higher than the faddict B feature of predetermined threshold is 0.2, have that its main orientation index equals or user's the ratio that is higher than the main orientation characteristic of predetermined threshold is 0.1, have that its connoisseur's index equals or the ratio of connoisseur's feature that is higher than the user of predetermined threshold is 0.02, and have that its most indexes equal or user's the ratio that is higher than most features of predetermined threshold is 0.1.

In addition, Figure 54 shows: project 2 is being made among the user of evaluation, have that its faddict A index equals or user's the ratio that is higher than the faddict A feature of predetermined threshold is 0, have that its faddict B index equals or user's the ratio that is higher than the faddict B feature of predetermined threshold is 0.03, have that its main orientation index equals or user's the ratio that is higher than the main orientation characteristic of predetermined threshold is 0.1, have that its connoisseur's index equals or user's the ratio that is higher than connoisseur's feature of predetermined threshold is 0.4, and have that its most indexes equal or user's the ratio that is higher than most features of predetermined threshold is 0.02.

Project statistical computation part 133 will be used to represent the information of owning rate of each user characteristics that the concern project is made the user of evaluation is offered item types determining section 134.

In step S472, whether the ratio of the evaluation that the user gave of feature that item types determining section 134 determines to have group 1 is very high.Specifically, item types determining section 134 obtains the concern project is made the owning rate sum of the user's of evaluation faddict A feature, faddict B feature and main orientation characteristic.If the owning rate sum that is obtained surpasses predetermined threshold, to determine to have the ratio of the evaluation that the user gave of feature of group 1 very high for item types determining section 134 so.After this, processing forwards step S473 to.

For example, according to Figure 53 and Figure 54, project 1 being made the user's of evaluation the faddict A feature, faddict B feature and the owning rate sum of main orientation characteristic is 0.6, and the owning rate sum of project 2 being made the user's of evaluation faddict A feature, faddict B feature and main orientation characteristic is 0.13.For example, if threshold value is set to 0.4, the ratio of evaluation that the user who determines to have the feature of group 1 so gives project 1 is very high, and the user who determines feature with group 1 to give the ratio of evaluation of project 2 not high.

In step S473,134 pairs of short-term hot topics of paying close attention to project of item types determining section are predicted.That is to say that 134 predictions of item types determining section will give repeatedly to estimate in the near future to the concern project.Item types determining section 134 will be used to represent that the information of the short-term hot topic of paying close attention to project having been predicted offers information and presents part 142.This information presents part 142 and will be the short-term hot topic be made prediction in the information of the concern project that such Statement of Facts kept to project information retaining part 143.After this, processing forwards step S474 to.

On the other hand, if the owning rate sum that is obtained in step S472 is equal to or less than predetermined threshold, to determine to have the ratio of the evaluation that the user gave of feature of group 1 not high for item types determining section 134 so, so the processing of skips steps S473, and handle and forward step S474 to.

In step S474, whether the ratio of the evaluation that the user gave of feature that item types determining section 134 determines to have group 2 is very high.Specifically, if the owning rate of connoisseur's feature of the user that the concern project is estimated surpasses predetermined threshold, to determine to have the ratio of the evaluation that the user gave of feature of group 2 very high for item types determining section 134 so.After this, processing forwards step S475 to.

For example, according to Figure 53 and Figure 54, the owning rate of connoisseur's feature of the user that project 1 is estimated is 0.02, and the owning rate of connoisseur's feature of the user who project 2 has been estimated is 0.4.For example, if threshold value is set to 0.3, the ratio of evaluation that the user who determines to have the feature of group 2 so gives project 1 is not high, and the user who determines to have the feature of group 2 to give the ratio of evaluation of project 2 very high.

In step S475,134 pairs of long-term hot topics of paying close attention to project of item types determining section are predicted.That is to say that 134 predictions of item types determining section are being estimated the concern project on the long duration very much.Item types determining section 134 will be used to represent that the information of the long-term hot topic of paying close attention to project having been predicted offers information and presents part 142.Information presents part 142 and will be long-term hot topic be made prediction in the information of the concern project that such Statement of Facts kept to project information retaining part 143.After this, processing forwards step S476 to.

On the other hand, if owning rate is equal to or less than predetermined threshold in step S474, to determine to have the ratio of the evaluation that the user gave of feature of group 2 not high for project category determining section 134 so, so the processing of skips steps S475, and handle and forward step S476 to.

In step S476, information presents part 142 user is presented in the hot topic prediction.For example, when the information that will pay close attention to project was presented to the user, information presented the information that part 142 also will be used to represent the hot topic prediction of this project and is sent to display part 122.Display part 122 shows the hot topic prediction and the information relevant with this project of concern project.For example, if the concern project is a music content, when predicting the short-term hot topic, show the message of similar " The hottest up and coming (the most fashionable arrival) " and so on so, and when predicting long-term hot topic, show the message of similar " Our pickupartist (artist that we select) " and so on.

Whether in this manner, estimate the project of can suitably predicting according to the user is popular.

Above-mentioned a series of processing can be carried out by hardware or by software.If this series of processes is carried out by software, the program that will constitute this software so is installed in the computing machine that is built in the specialized hardware or is installed to from program recorded medium for example can be by being installed to various programs in the general purpose personal computer of carrying out various functions in the general purpose personal computer.

Figure 55 has provided the example of hardware configuration that is used for carrying out by program the computing machine of above-mentioned a series of processing.

In this computing machine, CPU (CPU (central processing unit)) 301, ROM (ROM (read-only memory)) 302 and RAM (random access memory) 303 are connected with each other by bus 304.

Bus 304 also links to each other with input/output interface 305.Input/output interface 305 with disposed the importation 306 that forms by keyboard, mouse, microphone or the like, disposed the output 307 that forms, disposed the storage area 308 that forms, disposed the communications portion 309 that forms and the driver 310 that is used to drive such as the such removable media of disk, CD, magneto-optic disk or semiconductor memory links to each other by network interface or the like by hard disk, nonvolatile memory or the like by display or loudspeaker.

In the computing machine of configuration as mentioned above, for example, when CPU 301 is loaded among the RAM 303 by input/output interface 305 and bus 304 the program in the storage area 308 of will being stored in and carries out this program, carry out above-mentioned a series of processing.

The performed program of computing machine (CPU 301) is by it being recorded on the removable media 311 or by providing such as the so wired or wireless transmission medium of LAN (Local Area Network), internet or digital satellite broadcasting, wherein said removable media 311 is by disk (comprising floppy disk), CD (such as CD-ROM (compact disk-ROM (read-only memory)) or DVD (digital multi-purpose disk)), magneto-optic disk, semiconductor memory or the like the encapsulation medium that forms that disposes.

By removable media 311 is installed on the driver 310, can this program be installed in the storage area 308 by input/output interface 305.In addition, communications portion 309 can receive this program and attaches it in the storage area 308 by wired or wireless transmission medium.Otherwise, also this program can be installed in ROM 302 or the storage area 308 in advance.

The performed program of computing machine can be the order that occurs in this instructions according to them and carry out the program of processing in chronological order, but also can be parallel or such as call this program such in case of necessity between carry out the program of handling.

Employed term " system " is meant by a plurality of units or the like and is disposed the entire equipment that forms in this instructions.

In addition, embodiments of the invention are not limited to the foregoing description, and can make modification in all fields without departing from the scope of the invention.

Claims (19)

1. messaging device comprises:

The project evaluation deriving means is used to obtain the evaluation of estimate that each user gives each project;

User's statistical computation device, be used for calculating user's statistics of paying close attention to evaluation of user trend by use paying close attention to number of items that the user estimated, paying close attention to evaluation of estimate, each user that the user gives each project pay close attention at least one of evaluation of estimate of the project that the user estimated of the evaluation number of the project that the user estimated and each user of paying close attention; And

Present control device, be used for adding up to control the information relevant with project presented to paying close attention to the user according to the user.

2. according to the messaging device of claim 1, further comprise:

The project Extension arrangement is used for making the project cluster by preordering method,

Wherein user's statistical computation device calculates user's statistics according to the cluster specific distribution of paying close attention to the number of items that the user estimated.

3. according to the messaging device of claim 1, wherein:

User statistics comprises that group represent index, and on behalf of exponential representation, this group pay close attention to index of similarity between the cluster specific distribution of evaluation number of the affiliated whole group of the cluster specific distribution of the number of items that the user estimated and concern user.

4. according to the messaging device of claim 3, wherein:

User's statistics further comprises the modish index of representing the sequential mean value of index based on group.

5. according to the messaging device of claim 2, wherein:

User statistics comprises that index of conformity, this index of conformity pay close attention to the sequential stability index of the cluster specific distribution of the number of items that the user estimates.

6. according to the messaging device of claim 2, wherein:

User's statistics comprises the prejudice index, and this prejudice exponential representation is to the prejudice degree of the cluster specific distribution of the concern number of items that the user estimated.

7. according to any one messaging device in the claim 1 to 6, wherein:

Present control device and control, add up the information that represented concern user's feature is complementary so that select and present with the user to presenting.

8. according to any one messaging device in the claim 1 to 6, further comprise:

Project statistical computation device is used for the evaluation of estimate that gives according to each user and estimates at least one of number, calculates the project statistics of the evaluation trend that expression gives each project.

9. messaging device according to Claim 8, wherein:

User's statistical computation device according in the middle of the represented item characteristic of project statistics, pay close attention to the feature that bulk items had that the user estimated and calculate user's statistics of paying close attention to the user.

10. according to the messaging device of claim 9, wherein:

Project statistics comprises instantaneous index, in pass from mouth to mouth index and the designation number at least one, described instantaneous index is based on but underspeeding of the evaluation number of each each project estimated the relative value of the average velocity that number reduces with respect to become the time spent from each project, the described exponential representation of passing from mouth to mouth is to the length of the period of the evaluation number increase of each each project and the increase degree of estimating number, and described designation number is represented the sequential stability index to the evaluation number of each each project;

User's statistics comprises faddict's index, connoisseur's index, and in the conservative index at least one, described faddict's index based on estimated within the scheduled time slot that becomes in project after available and each ratio of project and the project that the user estimated of concern that has equaling or be higher than the instantaneous index of predetermined threshold, described connoisseur's index based on estimated within the scheduled time slot that becomes in project after available and each ratio of project and the project that the user estimated of concern that has equaling or be higher than the index that passes from mouth to mouth of predetermined threshold, described conservative index has equaling based on each or is higher than the ratio of project and the project that the user estimated of concern of the designation number of predetermined threshold.

11. according to the messaging device of claim 9, wherein:

Project statistics comprises fixedly fan's index of project, this project fixedly fan's index based on each user average ratings number to each each project within scheduled time slot; And

User statistics comprises fixedly fan's index of user, and this user is the fixedly project of fan's index and the ratio of the concern project that the user estimated of fan's index project of having equaling based on each or being higher than predetermined threshold fixedly.

12. messaging device according to Claim 8, wherein:

Project statistics comprises based on to the main index of the evaluation number of each each project and as the evaluation mean value of the mean value of the evaluation of estimate of each each project; And

User's statistics comprises faddict's index, main orientation index, general index number, and reputation orientation index, described faddict's index is based on the mean value of the main index of each each project that the user estimated of concern, described main orientation index is based on the correlativity of paying close attention between the main index that the user gives the evaluation of estimate of each each project and this project, described general index number is based on the correlativity between the evaluation mean value of paying close attention to evaluation of estimate that the user gives each each project and this project, and described reputation orientation index is based on the mean value of the evaluation mean value of concern each each project that the user estimated.

13. messaging device according to Claim 8, wherein:

Present control device make project statistics represented and add up the relevant highlighted demonstration of item characteristic of represented concern user's feature with the user and present.

14. messaging device according to Claim 8 further comprises:

Extraction element, be used to extract have project statistics represented and add up the project of the relevant feature of represented concern user's feature with the user,

Wherein present control device and control, so that the project of being extracted is presented to the concern user presenting.

15., further comprise according to any one messaging device in the claim 1 to 6:

User's index of similarity calculation element is used for adding up the user's index of similarity that calculates the index of similarity between the expression user according to the user;

The similar users extraction element is used to extract the similar users similar to paying close attention to the user; And

Extraction element is used to extract similar users and will recommends the project of paying close attention to the user to project conduct that it gives high evaluation value,

Wherein present control device and control, so that the project of being extracted is presented as the project that will recommend the concern user presenting.

16., further comprise according to any one messaging device in the claim 1 to 6:

User's index of similarity calculation element is used for adding up the user's index of similarity that calculates the index of similarity between the expression user according to the user;

The prediction and evaluation value calculation apparatus, be used for by using the pay close attention evaluation of estimate of project of other users, by distributing big weighting to the high evaluation of estimate that the user gave of itself and the value of user's index of similarity of paying close attention to the user, and pay close attention to the pay close attention prediction and evaluation value of project of user by distributing little weighting to the low evaluation of estimate that the user gave of itself and the value of user's index of similarity of paying close attention to the user, calculating; And

Extraction element is used to extract the high project of prediction and evaluation value as the project that will recommend the concern user,

Wherein present control device and control, so that the project of being extracted is presented as the project that will recommend the concern user presenting.

17. an information processing method that is used for messaging device may further comprise the steps:

Obtain the evaluation of estimate that each user gives each project;

By use paying close attention to number of items that the user estimated, paying close attention to evaluation of estimate, each user that the user gives each project pay close attention in the evaluation of estimate of the project that the user estimated at least one of the evaluation number of the project that the user estimated and each user of paying close attention, calculate user's statistics of paying close attention to evaluation of user trend; And

Add up to control the information relevant with project presented to according to the user and pay close attention to the user.

18. one kind is used to make computing machine to carry out the program of the processing that may further comprise the steps:

Obtain the evaluation of estimate that each user gives each project;

By use paying close attention to number of items that the user estimated, paying close attention to evaluation of estimate, each user that the user gives each project pay close attention in the evaluation of estimate of the project that the user estimated at least one of the evaluation number of the project that the user estimated and each user of paying close attention, calculate user's statistics of paying close attention to evaluation of user trend; And

Add up to control the information relevant with project presented to according to the user and pay close attention to the user.

19. a messaging device comprises:

Part is obtained in project evaluation, is configured to obtain the evaluation of estimate that each user gives each project;

User's statistical computation part, be configured to calculate user's statistics of paying close attention to evaluation of user trend by use paying close attention to number of items that the user estimated, paying close attention to evaluation of estimate, each user that the user gives each project pay close attention in the evaluation of estimate of the project that the user estimated at least one of the evaluation number of the project that the user estimated and each user of paying close attention; And

Present control section, be configured to add up to control the information relevant with project presented to pay close attention to the user according to the user.

Images