CN105653735A - Network resource recommendation method and device - Google Patents
Network resource recommendation method and device Download PDFInfo
- Publication number
- CN105653735A CN105653735A CN201610112975.2A CN201610112975A CN105653735A CN 105653735 A CN105653735 A CN 105653735A CN 201610112975 A CN201610112975 A CN 201610112975A CN 105653735 A CN105653735 A CN 105653735A
- Authority
- CN
- China
- Prior art keywords
- user
- network resource
- moment
- recommendation
- target network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention relates to a network resource recommendation method and device. The network resource recommendation method comprises the steps that network resource access requests of first users are received; the identification information of second users accessing target network resources are obtained from behavior data of the first users according to the target network resources; the historical network resources accessed by the second users are obtained from the pre-collected behavior data of the second users according to the identification information of the second users; first recommendation lists of the target network resources are returned to the first users based on the correlation between the target network resources and the historical network resources.
Description
Technical field
The present invention relates to Internet resources process field, particularly relate to a kind of network resource recommended method and device.
Background technology
Along with developing rapidly of Internet technology, Internet resources quantity constantly increases, content is also enriched constantly. But this brings many challenges to the network user, they have to spend substantial amounts of time and efforts, just can filter out oneself desired Internet resources. Therefore, the commending system filtering out the Internet resources that each user expectation obtains from substantial amounts of Internet resources becomes the main tool solving this problem. These commending systems utilize the historical behavior data of user, carry out personalized calculating, it has been found that user interest point, experience thus recommending user to be likely to content interested, the time largely saving user the online provided to user.
It is presently recommended that the algorithm that system is commonly used has collaborative filtering (CollaborativeFiltering, CF), clicking rate (Click-throughRate, CTR) to estimate. For the Internet resources that short-time exposure amount is high or history access record is abundant, these algorithms most in use can obtain comparing high-quality, stable recommendation results. But, owing to these algorithms are dependent on the accumulation (daily record produced for user mainly arranges in units of day) of the historical data of long period, therefore playing, for the Internet resources (such as new uploaded videos, audio frequency) newly uploaded or history, the content recommendation that the relatively low old video of behavior, audio frequency etc. obtain, these algorithms need to wait for a long time or just can produce to compare the content recommendation of high-quality after substantial amounts of user's visit capacity.
In view of this problem, currently occurring in that some follow-on quasi real time CF algorithms, before it allows to conduct interviews based on user, the history of 6 hours, 12 hours, one day etc. accesses information and carries out the recommendation of Internet resources, thus solving above-mentioned technical problem. But there is following defect in these algorithms.
Firstly, since these algorithms need to carry out the collection of historical information, analysis before carrying out the calculating of recommendation results every time, ratio is relatively time-consuming. Secondly as these quasi real time have to be larger than it and can provide the time period needed for recommendation results in the allocating time interval of CF algorithm, its calculate problem consuming time result in again frequency that these algorithms call can not be too high.Therefore, when user carries out network resource accession, there is the situation not yet completing due to quasi real time CF algorithm to calculate and cannot providing recommendation results, this will cause not good enough Consumer's Experience to the network user.
Summary of the invention
Technical problem
In consideration of it, the technical problem to be solved in the present invention is, how quickly to provide the recommendation results of Internet resources, it is provided that good Consumer's Experience.
Technical scheme
In order to solve above-mentioned technical problem, the present invention provides a kind of recommendation method of Internet resources, comprising:
Receiving the network resource accession request of first user, the request of described network resource accession includes the target network resource that the request of described first user accesses, and described first user is the user initiating the request of described network resource accession;
From first user behavioral data, the identification information of each second user accessing described target network resource is obtained according to described target network resource, described first user behavioral data include collected from the first moment to the second moment each described second user access the data of described target network resource, before being engraved in described second moment when described first;
Identification information according to each described second user, obtaining each historical network resource that each described second user accesses from the second user behavior data collected in advance, described second user behavior data includes the data of the Internet resources that collected each described second user before described first moment accesses;
Based on the dependency between described target network resource and each described historical network resource, return the first recommendation list of described target network resource to described first user.
For said method, in a kind of possible implementation, after receiving the network resource accession request of first user, also include:
Judge whether to utilize the second recommendation list that quasi real time collaborative filtering recommending method produces;
When there is described second recommendation list, return described second recommendation list to described first user;
When being absent from described second recommendation list, perform to obtain from first user behavioral data the step of the identification information of each second user accessing described target network resource according to described target network resource.
For said method, in a kind of possible implementation, after receiving the network resource accession request of first user, also include:
Judge whether content recommendation;
When there is described content recommendation, perform the step judging whether to utilize quasi real time the second recommendation list that collaborative filtering recommending method produces;
When being absent from described content recommendation, generate the 3rd recommendation list according to hot spot networks resource, and return described 3rd recommendation list to described first user.
For said method, in a kind of possible implementation, before receiving the network resource accession request of first user, also include:
History data collection step, described first moment collect described first moment for the previous period in the data of Internet resources that access of each user;
Current data collection step, collects the data of the Internet resources that each user between described first moment and described second moment accesses in described second moment.
For said method, in a kind of possible implementation, being historical time before described first moment, be time of day between described first moment to described second moment, described recommendation method also includes: perform once described Current data collection step every the interval set.
For said method, in a kind of possible implementation, also include: will add in described first recommendation list at the historical network resource setting more than threshold value with the dependency of described target network resource.
The present invention also provides for the recommendation apparatus of a kind of Internet resources, comprising:
Receiver module, for receiving the network resource accession request of first user, the request of described network resource accession includes the target network resource that the request of described first user accesses, and described first user is the user initiating the request of described network resource accession;
User profile acquisition module, it is connected with described receiver module, for obtaining the identification information of each second user accessing described target network resource from first user behavioral data according to described target network resource, described first user behavioral data include collected from the first moment to the second moment each described second user access the data of described target network resource, before being engraved in described second moment when described first;
History source obtaining module, it is connected with described user profile acquisition module, for the identification information according to each described second user, obtaining each historical network resource that each described second user accesses from the second user behavior data collected in advance, described second user behavior data includes the data of the Internet resources that collected each described second user before described first moment accesses;
Recommending module, is connected with described receiver module and described history source obtaining module, for based on the dependency between described target network resource and each described historical network resource, returning the first recommendation list of described target network resource to described first user.
For said apparatus, in a kind of possible implementation, also include:
First judge module, it is connected respectively with described receiver module and described user profile acquisition module, after network resource accession for receiving first user at described receiver module is asked, it may be judged whether exist and utilize the second recommendation list that quasi real time collaborative filtering recommending method produces; When there is described second recommendation list, return described second recommendation list to described first user;
Described user profile acquisition module is additionally operable to, when described first judge module judges to be absent from described second recommendation list, from first user behavioral data, obtain the identification information of each second user accessing described target network resource according to described target network resource.
For said apparatus, in a kind of possible implementation, also include:
Second judge module, is connected respectively with described receiver module and described first judge module, after the network resource accession for receiving first user at described receiver module is asked, it may be judged whether there is content recommendation; When being absent from described content recommendation, generate the 3rd recommendation list according to hot spot networks resource, and return described 3rd recommendation list to described first user;
Described first judge module is additionally operable to when described second judge module judges there is described content recommendation, it may be judged whether exists and utilizes the second recommendation list that quasi real time collaborative filtering recommending method produces.
For said apparatus, in a kind of possible implementation, also include:
History data collection module, for described first moment collect described first moment for the previous period in the data of Internet resources that access of each user;
Current data collection module, for collecting the data of the Internet resources that each user between described first moment and described second moment accesses in described second moment.
For said apparatus, in a kind of possible implementation, it was historical time before described first moment, being time of day between described first moment to described second moment, described Current data collection module is additionally operable to perform once to collect the step of data of the Internet resources that each user between described first moment and described second moment accesses in described second moment every the interval set.
For said apparatus, in a kind of possible implementation, described recommending module is additionally operable to add in described first recommendation list at the historical network resource setting more than threshold value with the dependency of described target network resource.
Beneficial effect
The network resource recommended method of the present invention, information is accessed by calculated in advance history before conducting interviews user, therefore compare quasi real time CF algorithm and can save the calculating time of recommendation results, and can relatively frequently call soon owing to calculating speed, therefore for being in the Internet resources of cold start-up, it is not required that after a long or could obtain after substantial amounts of user's visit capacity correspondence recommendation results. It addition, the recommendation method of the present invention can serve as off-line CF algorithm and quasi real time CF algorithm supplement.
By below with reference to the accompanying drawing explanation to exemplary embodiments, the further feature of the present invention will be apparent from.
Accompanying drawing explanation
Fig. 1 illustrates the flow chart of network resource recommended method according to an embodiment of the invention.
Fig. 2 schematically shows the flow chart of network resource recommended method according to another embodiment of the present invention.
Fig. 3 A illustrates the example of the accessing request information to the Internet resources to access received in network resource recommended method according to another embodiment of the present invention.
Fig. 3 B illustrates the example of the user's behavioral data on the same day in network resource recommended method according to another embodiment of the present invention.
Fig. 3 C illustrates the example of the user's historical behavior data in network resource recommended method according to another embodiment of the present invention.
Fig. 4 A and Fig. 4 B is shown respectively the application responding process figure according to the network resource request before and after the network resource recommended method of one embodiment of the invention.
Fig. 5 illustrates the structural representation of the recommendation apparatus of Internet resources according to an embodiment of the invention.
Fig. 6 illustrates the structural representation of the recommendation apparatus of Internet resources according to another embodiment of the present invention.
Detailed description of the invention
The various exemplary embodiments of the present invention, feature and aspect is described in detail below with reference to accompanying drawing. Accompanying drawing labelling identical in accompanying drawing represents the same or analogous element of function. Although the various aspects of embodiment shown in the drawings, but unless otherwise indicated, it is not necessary to accompanying drawing drawn to scale.
Word " exemplary " special here means " as example, embodiment or illustrative ". Here should not necessarily be construed as preferred or advantageous over other embodiments as any embodiment illustrated by " exemplary ".
It addition, in order to better illustrate the present invention, detailed description of the invention below gives numerous details. It will be appreciated by those skilled in the art that there is no some detail, the equally possible enforcement of the present invention. In some instances, method, means, element and the circuit known for those skilled in the art are not described in detail, in order to highlight the purport of the present invention.
Embodiment 1
Fig. 1 illustrates the flow chart of network resource recommended method according to an embodiment of the invention, as it is shown in figure 1, this network resource recommended method mainly may comprise steps of:
Step 100, the reception first user access request to Internet resources.
Here Internet resources refer to the summation by the utilizable various information resources of computer network. Specifically, can refer to all with spreadsheet format, the information (such as the multimedia resource such as video, audio frequency and document etc.) of the various ways such as word, image, sound, animation is stored in the carrier of the non-paper medium such as light, magnetic, and by resource that the modes such as network service, computer or terminal reproduce out.Here first user is the user initiating the access request to Internet resources. User asks mark (identification, the ID) information of the Internet resources accessed at least including this user in the access request of Internet resources, and an id information generally can uniquely identify certain Internet resources. This access request can also include the data such as the classification information of these Internet resources, label information, title.
Fig. 3 A illustrates the example of the accessing request information to Internet resources (hereinafter referred to as target network resource) received in network resource recommended method according to an embodiment of the invention from user.
Internet resources ID301 is the information for uniquely identifying Internet resources, the classification 302 of Internet resources and label 303 are the important evidence of search target network resource, the title 304 of Internet resources can be the brief and concise summary to these Internet resources, it is also possible to be network resource provider's name to it. Such as, receive and include information: Internet resources ID is " AAA0001 ", be categorized as " imperial palace is acute ", label is " grandson pari ", title is the accessing request information of " the Mi month passes ". Accessing request information shown in Fig. 3 A every exemplary only, its every and corresponding value comprised is not limited to this.
Step 101, from first user behavioral data, obtain the identification information of each second user accessing described target network resource according to described target network resource.
First user behavioral data can be such as user's behavioral data on the same day. User's behavioral data on the same day is collected to access data from produced by customer access network resource the second moment of the first moment to the same day on the same day, wherein first time be engraved in for the second moment before. User's behavioral data on the same day at least includes the identification information of target network resource and the identification information (as the cookie data of the user conducted interviews) of the user that have accessed these Internet resources. User's behavioral data on the same day can also include the attribute information etc. of the Internet resources such as the title of the classification of such as Internet resources, the label of Internet resources, Internet resources.
Fig. 3 B illustrates the example of the user's behavioral data on the same day in network resource recommended method according to another embodiment of the present invention. Label identical for Fig. 3 B and Fig. 3 A has identical implication, therefore no longer repeats one by one. ID 305 in user's behavioral data on the same day is for identifying the user from the second moment of the first moment to the same day on the same day, these Internet resources accessed.
Such as, when receiving the access request of target network resource that Internet resources ID is " AAA0001 ", scan for from the user's behavioral data on the same day shown in Fig. 3 B, thus obtaining corresponding item 1001,1002. Using the identification information as each user accessing this target network resource from the second moment of the first moment to the same day on the same day of ID U1, the U2 in item 1001,1002.
Step 102, identification information according to each second user obtained in step 101 obtain, from the second user behavior data collected in advance, the historical network resource that these users were accessed.
Second user behavior data can be the historical behavior data in such as user nearest a period of time. User's historical behavior data are that in a period of time before the first moment, history produced by customer access network resource accesses network resource information. The identification information of the Internet resources that user's historical behavior data can include the user totem information of such as user cookie etc. and these users had accessed.User's historical behavior data can also include accessing the data such as duration, user interaction data, Classification of Web Resources, Internet resources label and Internet resources quality.
Fig. 3 C illustrates the example of the user's historical behavior data in network resource recommended method according to another embodiment of the present invention. Label identical for Fig. 3 C and Fig. 3 A, 3B has identical implication, therefore no longer repeats one by one. Access duration 306 in user's historical behavior data represents that user is in accessed Internet resources place page residence time, belongs to the behavioral data of user. The behavioral data of user can also include such as whether dragged progress bar, whether convergent-divergent crosses the interactive datas such as window. Different types of Internet resources are represented different contents by Internet resources quality 307, for instance for video being high definition, SD, super clear etc., for audio frequency can be lossless, damage.
Obtain accessing the ID U1 of target network resource " AAA0001 ", U2 from the second moment of the first moment to the same day on the same day in a step 101, scan for from the user's historical behavior data shown in Fig. 3 C, thus obtaining corresponding item 1003,1004,1005,1006. Using the Internet resources in item 1003,1004,1005,1006 as the historical network resource accessed in a period of time before the first moment of these users.
Step 103, dependency between based target Internet resources and historical network resource determine the first recommendation list of target network resource, and return the first recommendation list of described target network resource to first user.
Specifically, integrated treatment and filtration can be carried out according to information such as the user behavior (as accessed duration etc.) of the classification of visited network resource, label and each historical network resource, Classification of Web Resources, Internet resources label, Internet resources quality, obtain the Internet resources that target network resource dependency is the highest, obtain its TOP content recommendation list (example of the first recommendation list) as this target network resource. Such as, respective weight is given respectively by the user behavior of each historical network resource, Classification of Web Resources, Internet resources label, Internet resources quality etc., calculate the summation of its every weighted value for each historical network resource, and the summation of the weighted value of each historical network resource is ranked up from high to low. Concurrently set threshold value, weighted value being added recommendation list higher than the Internet resources of threshold value, not considering lower than the Internet resources of threshold value, thus obtaining the recommendation list of visited network resource.
But, the determination of the dependency between target network resource and historical network resource is not limited to said method.
Network resource recommended method according to the present embodiment calculates the recommendation list of target network resource owing to using the second user behavior data collected in advance, rather than collect temporarily and arrange user's historical behavior data when calculating recommendation list, therefore consuming time few for quasi real time CF algorithm, and support high-frequency calling.
Embodiment 2
In this embodiment of the invention, and differring primarily in that of embodiment 1, also include user's history data collection step, for the first moment collect the first moment for the previous period in the data of Internet resources of user's access; And user's behavioral data on same day collection step, for collecting the data in the first moment and the Internet resources of the accesses of users between in the second moment in the second moment.
Fig. 2 schematically shows the flow chart of network resource recommended method according to another embodiment of the present invention. Label identical for Fig. 2 and Fig. 1 has identical content, and with differring primarily in that of Fig. 1, this recommendation method can also comprise the following steps before step 100:
Step 201, it is judged that current time.
The judgement of current time such as can start every 1 hour, 1.5 hours or 2 hours etc. from the moment in morning (example in the first moment) and carry out once. For example, it is contemplated that the visit capacity to festivals or holidays such as weekends is likely to increase to some extent relative to workaday visit capacity, therefore, it can be set to the judgement interval of the festivals or holidays such as weekend 1 hour, and workaday judgement interval is set to 1.5 hours. Furthermore, it is contemplated that morning��8 front visit capacities are relatively fewer, 8 initially enter access rush hour, therefore can the judgement of current time be set as from 8 a.m..
Step 202, it may be judged whether be new one day. When step 202 be judged as be not new one, enter step 203. When step 202 is judged as YES new one, enter step 204.
Step 203, collects the data of the Internet resources that the user between the first moment and the second moment before current time accesses. Afterwards, step 205 is entered.
Step 204, collects the data of the Internet resources of interior for the previous period user's access in the first moment, obtains the data after simplifying in a corresponding format. Here a period of time can be 24 hours, 36 hours, 48 hours etc. Afterwards, it is back to step 202. The data of the Internet resources that user accesses include such as ID, network resource identifier, User IP, access timestamp, access the information such as duration, accession page, device type. Data after simplification such as can include ID, network resource identifier etc. Further, it is possible to the convergence after simplifying is obtained the Internet resources list such as list of videos that each user's history accesses, and each video can also comprise essential information or the relevant information of some videos here.
Step 205, when receiving user to the access request of Internet resources, enters step 101.
Network resource recommended method according to the present embodiment is by collecting user's historical behavior data in advance in times such as user's visit capacity less mornings, not only reduce calculate recommendation list time consuming time, fight for Internet resources without in peak period with user simultaneously, reduce network burden. Due to consuming time few, therefore support calling of upper frequency. It addition, existing CF algorithm generally carries out the collection of data in units of day, therefore for the less Internet resources of visit capacity, it is impossible to obtain the content recommendation of high-quality or be unsatisfactory for calculating demand and content recommendation cannot be obtained. Even if the off-line CF that CF algorithm is optimized and quasi real time CF algorithm etc., but all need the data of long period and calculating consuming time, this meeting is less popular for some or cannot accumulate the Internet resources of certain visit capacity in the short time, is difficult to quickly obtain optimum content recommendation and maybe cannot obtain optimum content recommendation. And the network resource recommended method according to the present embodiment is from another angle, recommendation results to be optimized, can combine with such as off-line CF, quasi real time CF algorithm etc., before above-mentioned CF algorithm does not come into force or when cannot come into force, it is possible to obtain better recommendation results.Such as, when carrying out the collection of user's behavioral data on the same day 8 o'clock (example in the second moment), within 8: 05, new Internet resources are had to upload and when the access information of correspondence, when 9 o'clock, (example in the second moment) carried out the collection of user's behavioral data on the same day, the access information of these new Internet resources will be searched such that it is able to obtain the recommendation list of these new Internet resources.
Embodiment 3
Fig. 4 A and 4B is shown respectively the application responding process figure according to the network resource request before and after the network resource recommended method (hereinafter referred to as supplementary recommendation method) of one embodiment of the invention.
As shown in Figure 4 A, when not applying network resource recommended method according to an embodiment of the invention, network resource accession for user is asked, the recommendation list (such as, the second recommendation list) passing back through the quasi real time target network resource that CF algorithm is calculated is responded. This responding process comprises the steps:
Step 401, receives the request of customer access network resource. This step is identical with the step 205 in embodiment 2, but can not directly perform step 101 after step 401, and is carried out step 402.
Step 402, it may be judged whether there is content recommendation. If there is content recommendation, then enter step 403, if there is no content recommendation, then enter step 404.
In general, quasi real time the result of CF algorithm and supplementary recommendation method according to embodiments of the present invention all can combine in a manner that recommendation list and be stored, and responds for the request for customer access network resource. When user asks to access Internet resources, it is judged that whether these Internet resources exist recommendation list, if it is present think and there is content recommendation. If it does not exist, then think and be absent from content recommendation. It addition, the consequently recommended list of storage on line periodically will be updated. Such as, when quasi real time CF algorithm not yet calculates recommendation list, supplementary recommendation method according to embodiments of the present invention is utilized to calculate recommendation list and store, and after quasi real time CF algorithm calculates recommendation list, by utilizing the recommendation list that quasi real time CF algorithm calculates, the recommendation list of final storage is updated.
Step 403, utilizes quasi real time CF algorithm to calculate the recommendation network resource the second recommendation list as target network resource. Enter step 405 afterwards.
Step 404, obtains the hot spot networks resource the 3rd recommendation list as target network resource of calculating. Enter step 405 afterwards.
Step 405, returns calculated recommendation list.
As shown in Figure 4 B, the network resource accession for user is asked, and responds applying above-mentioned supplementary recommendation method before not yet coming into force at quasi real time CF algorithm to the recommendation list calculating target network resource.
Number of steps identical for Fig. 4 B and 4A has identical content, and with differring primarily in that of Fig. 4 A, this responding process can also include step 406 and step 407.
Step 406, it may be judged whether exist and utilize the quasi real time calculated recommendation list of CF algorithm. If it is present enter step 403. If it does not exist, then enter step 407.
Step 407, utilizes the supplementary recommendation method in the step 101-103 in Fig. 1 to calculate recommendation list.
Afterwards, enter step 405, calculated recommendation list is returned as recommendation results.
When quasi real time CF algorithm not yet calculates the recommendation list of target network resource, it is possible to provide, as the supplementary of quasi real time CF algorithm, the content recommendation comparing high-quality by the supplementary recommendation method of recommendation list can be calculated at short notice.
Embodiment 4
Fig. 5 illustrates the structural representation of the recommendation apparatus of Internet resources according to an embodiment of the invention. The recommendation apparatus of these Internet resources may include that receiver module 41, user profile acquisition module 42, history source obtaining module 43 and recommending module 44.
Receiver module 41 is for receiving the network resource accession request of first user, and the request of described network resource accession includes the target network resource that the request of described first user accesses, and described first user is the user initiating the request of described network resource accession.
User profile acquisition module 42 is connected with described receiver module 41, for obtaining the identification information of each second user accessing described target network resource from first user behavioral data according to described target network resource, described first user behavioral data include collected from the first moment to the second moment each described second user access the data of described target network resource, before being engraved in described second moment when described first.
History source obtaining module 43 is connected with described user profile acquisition module 42, for the identification information according to each described second user, obtaining each historical network resource that each described second user accesses from the second user behavior data collected in advance, described second user behavior data includes the data of the Internet resources that collected each described second user before described first moment accesses.
Recommending module 44 is connected with described receiver module 41 and described history source obtaining module 43, for based on the dependency between described target network resource and each described historical network resource, returning the first recommendation list of described target network resource to described first user.
Specifically, the receiver module 41 of the recommendation apparatus of the Internet resources of the present embodiment, user profile acquisition module 42, history source obtaining module 43 and recommending module 44, it is possible to the recommendation method of the Internet resources in execution above-described embodiment. Concrete principle and example may refer to the associated description in above-described embodiment.
In the present embodiment, from the second user behavior data collected in advance, obtain, by history source obtaining module 43, each historical network resource that each described second user accesses and can be substantially reduced and calculate the consuming time of recommendation list, and support high-frequency calling.
Embodiment 5
Fig. 6 illustrates the structural representation of the recommendation apparatus of Internet resources according to another embodiment of the present invention. Being distinctive in that with a upper embodiment, this network resource recommended device also includes with lower module:
History data collection module 45, for described first moment collect described first moment for the previous period in the data of Internet resources that access of each user.
Current data collection module 46, for collecting the data of the Internet resources that each user between described first moment and described second moment accesses in described second moment.
In a kind of possible implementation, this network resource recommended device also includes the first judge module 47, it is connected respectively with described receiver module 41 and described user profile acquisition module 42, and the network resource accession for receiving first user at described receiver module 41 judges whether to utilize the second recommendation list that quasi real time collaborative filtering recommending method produces after asking. When there is described second recommendation list, return described second recommendation list to described first user. User profile acquisition module 42 is additionally operable to, when described first judge module 47 judges to be absent from described second recommendation list, obtain the identification information of each second user accessing described target network resource from first user behavioral data according to described target network resource.
In a kind of possible implementation, this network resource recommended device also includes the second judge module 48, it is connected respectively with described receiver module 41 and described first judge module 47, after network resource accession for receiving first user at described receiver module 41 is asked, it may be judged whether there is content recommendation. When being absent from described content recommendation, generate the 3rd recommendation list according to hot spot networks resource, and return described 3rd recommendation list to described first user. Described first judge module 47 is additionally operable to when described second judge module 48 judges there is described content recommendation, it may be judged whether exists and utilizes the second recommendation list that quasi real time collaborative filtering recommending method produces.
In a kind of possible implementation, it was historical time before described first moment, being time of day between described first moment to described second moment, described Current data collection module 46 is additionally operable to perform once to collect the step of data of the Internet resources that each user between described first moment and described second moment accesses in described second moment every the interval set.
In a kind of possible implementation, described recommending module 44 is additionally operable to add in described first recommendation list at the historical network resource setting more than threshold value with the dependency of described target network resource.
The recommendation apparatus of the Internet resources of the present embodiment, when quasi real time CF algorithm not yet calculates the recommendation list of target network resource, it is possible to provide, as the supplementary of quasi real time CF algorithm, the content recommendation comparing high-quality by the supplementary recommendation method of recommendation list can be calculated at short notice.
The above; being only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, any those familiar with the art is in the technical scope that the invention discloses; change can be readily occurred in or replace, all should be encompassed within protection scope of the present invention. Therefore, protection scope of the present invention should be as the criterion with described scope of the claims.
Claims (12)
1. the recommendation method of Internet resources, it is characterised in that including:
Receiving the network resource accession request of first user, the request of described network resource accession includes the target network resource that the request of described first user accesses, and described first user is the user initiating the request of described network resource accession;
From first user behavioral data, the identification information of each second user accessing described target network resource is obtained according to described target network resource, described first user behavioral data include collected from the first moment to the second moment each described second user access the data of described target network resource, before being engraved in described second moment when described first;
Identification information according to each described second user, obtaining each historical network resource that each described second user accesses from the second user behavior data collected in advance, described second user behavior data includes the data of the Internet resources that collected each described second user before described first moment accesses;
Based on the dependency between described target network resource and each described historical network resource, return the first recommendation list of described target network resource to described first user.
2. recommendation method according to claim 1, it is characterised in that after receiving the network resource accession request of first user, also include:
Judge whether to utilize the second recommendation list that quasi real time collaborative filtering recommending method produces;
When there is described second recommendation list, return described second recommendation list to described first user;
When being absent from described second recommendation list, perform to obtain from first user behavioral data the step of the identification information of each second user accessing described target network resource according to described target network resource.
3. recommendation method according to claim 2, it is characterised in that after receiving the network resource accession request of first user, also include:
Judge whether content recommendation;
When there is described content recommendation, perform the step judging whether to utilize quasi real time the second recommendation list that collaborative filtering recommending method produces;
When being absent from described content recommendation, generate the 3rd recommendation list according to hot spot networks resource, and return described 3rd recommendation list to described first user.
4. recommendation method according to any one of claim 1 to 3, it is characterised in that before receiving the network resource accession request of first user, also include:
History data collection step, described first moment collect described first moment for the previous period in the data of Internet resources that access of each user;
Current data collection step, collects the data of the Internet resources that each user between described first moment and described second moment accesses in described second moment.
5. recommendation method according to claim 4, it is characterized in that, being historical time before described first moment, be time of day between described first moment to described second moment, described recommendation method also includes: perform once described Current data collection step every the interval set.
6. recommendation method according to any one of claim 1 to 5, it is characterised in that also include: will add in described first recommendation list at the historical network resource setting more than threshold value with the dependency of described target network resource.
7. the recommendation apparatus of Internet resources, it is characterised in that including:
Receiver module, for receiving the network resource accession request of first user, the request of described network resource accession includes the target network resource that the request of described first user accesses, and described first user is the user initiating the request of described network resource accession;
User profile acquisition module, it is connected with described receiver module, for obtaining the identification information of each second user accessing described target network resource from first user behavioral data according to described target network resource, described first user behavioral data include collected from the first moment to the second moment each described second user access the data of described target network resource, before being engraved in described second moment when described first;
History source obtaining module, it is connected with described user profile acquisition module, for the identification information according to each described second user, obtaining each historical network resource that each described second user accesses from the second user behavior data collected in advance, described second user behavior data includes the data of the Internet resources that collected each described second user before described first moment accesses;
Recommending module, is connected with described receiver module and described history source obtaining module, for based on the dependency between described target network resource and each described historical network resource, returning the first recommendation list of described target network resource to described first user.
8. recommendation apparatus according to claim 7, it is characterised in that also include:
First judge module, it is connected respectively with described receiver module and described user profile acquisition module, after network resource accession for receiving first user at described receiver module is asked, it may be judged whether exist and utilize the second recommendation list that quasi real time collaborative filtering recommending method produces;When there is described second recommendation list, return described second recommendation list to described first user;
Described user profile acquisition module is additionally operable to, when described first judge module judges to be absent from described second recommendation list, from first user behavioral data, obtain the identification information of each second user accessing described target network resource according to described target network resource.
9. recommendation apparatus according to claim 8, it is characterised in that also include:
Second judge module, is connected respectively with described receiver module and described first judge module, after the network resource accession for receiving first user at described receiver module is asked, it may be judged whether there is content recommendation; When being absent from described content recommendation, generate the 3rd recommendation list according to hot spot networks resource, and return described 3rd recommendation list to described first user;
Described first judge module is additionally operable to when described second judge module judges there is described content recommendation, it may be judged whether exists and utilizes the second recommendation list that quasi real time collaborative filtering recommending method produces.
10. the recommendation apparatus according to any one of claim 7 to 9, it is characterised in that also include:
History data collection module, for described first moment collect described first moment for the previous period in the data of Internet resources that access of each user;
Current data collection module, for collecting the data of the Internet resources that each user between described first moment and described second moment accesses in described second moment.
11. recommendation apparatus according to claim 10, it is characterized in that, it was historical time before described first moment, being time of day between described first moment to described second moment, described Current data collection module is additionally operable to perform once to collect the step of data of the Internet resources that each user between described first moment and described second moment accesses in described second moment every the interval set.
12. the recommendation apparatus according to any one of claim 7 to 11, it is characterised in that described recommending module is additionally operable to add in described first recommendation list at the historical network resource setting more than threshold value with the dependency of described target network resource.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610112975.2A CN105653735A (en) | 2016-02-29 | 2016-02-29 | Network resource recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610112975.2A CN105653735A (en) | 2016-02-29 | 2016-02-29 | Network resource recommendation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105653735A true CN105653735A (en) | 2016-06-08 |
Family
ID=56492919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610112975.2A Pending CN105653735A (en) | 2016-02-29 | 2016-02-29 | Network resource recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105653735A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106303720A (en) * | 2016-08-02 | 2017-01-04 | 合网络技术（北京）有限公司 | A kind of video recommendation method and system |
| CN106603296A (en) * | 2016-12-20 | 2017-04-26 | 北京奇虎科技有限公司 | Log processing method and device |
| CN109002530A (en) * | 2018-07-17 | 2018-12-14 | 佛山市灏金赢科技有限公司 | A kind of resource recommendation method and device |
| CN110020176A (en) * | 2017-12-29 | 2019-07-16 | ***通信集团公司 | A kind of resource recommendation method, electronic equipment and computer readable storage medium |
| CN111277898A (en) * | 2018-12-05 | 2020-06-12 | ***通信集团广西有限公司 | Content pushing method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101916286A (en) * | 2010-08-23 | 2010-12-15 | 宇龙计算机通信科技(深圳)有限公司 | Information recommendation method and system |
| CN101957834A (en) * | 2010-08-12 | 2011-01-26 | 百度在线网络技术（北京）有限公司 | Content recommending method and device based on user characteristics |
| US20140344254A1 (en) * | 2011-12-14 | 2014-11-20 | Beijing Qihood Technology Company Limited | Software recommending method and recommending system |
-
2016
- 2016-02-29 CN CN201610112975.2A patent/CN105653735A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101957834A (en) * | 2010-08-12 | 2011-01-26 | 百度在线网络技术（北京）有限公司 | Content recommending method and device based on user characteristics |
| CN101916286A (en) * | 2010-08-23 | 2010-12-15 | 宇龙计算机通信科技(深圳)有限公司 | Information recommendation method and system |
| US20140344254A1 (en) * | 2011-12-14 | 2014-11-20 | Beijing Qihood Technology Company Limited | Software recommending method and recommending system |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106303720A (en) * | 2016-08-02 | 2017-01-04 | 合网络技术（北京）有限公司 | A kind of video recommendation method and system |
| CN106603296A (en) * | 2016-12-20 | 2017-04-26 | 北京奇虎科技有限公司 | Log processing method and device |
| CN110020176A (en) * | 2017-12-29 | 2019-07-16 | ***通信集团公司 | A kind of resource recommendation method, electronic equipment and computer readable storage medium |
| CN109002530A (en) * | 2018-07-17 | 2018-12-14 | 佛山市灏金赢科技有限公司 | A kind of resource recommendation method and device |
| CN111277898A (en) * | 2018-12-05 | 2020-06-12 | ***通信集团广西有限公司 | Content pushing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11200596B2 (en) | System and method for segmenting and targeting audience members | |
| CN100517293C (en) | Method for extracting content, content extraction server based on RSS and apparatus for managing the same and system for providing standby screen of mobile communication terminal using the same | |
| CN105653735A (en) | Network resource recommendation method and device | |
| KR101385700B1 (en) | Method and apparatus for providing moving image advertisements | |
| US8819716B2 (en) | Apparatus, system, method and computer readable recording medium storing the program for related recommendation of TV program contents and web contents | |
| WO2006052444A2 (en) | Audience targeting system with profile synchronization | |
| WO2006052442A2 (en) | Audience targeting system with segment management | |
| CN109151824B (en) | Library data service expansion system and method based on 5G architecture | |
| CN105183873A (en) | Malicious clicking behavior detection method and device | |
| CN104077714A (en) | Method and system for acquiring preference of user visiting website and pushing advertisements to user visiting website | |
| CN111597449B (en) | Candidate word construction method and device for search, electronic equipment and readable medium | |
| CN103324696B (en) | A kind of data log collection and statistical analysis system and method | |
| EP2764449A1 (en) | Method and apparatuses for enabling recommendations | |
| CN102710773A (en) | Method and system for recommending network information | |
| CN101140582A (en) | User preferences generation system and method thereof | |
| WO2009116723A1 (en) | Method and system for providing url possible new advertising | |
| CN104077713A (en) | Method and system of demand assessment and advertisement push for user visiting website | |
| CN106603351B (en) | Television advertisement pushing method and system based on user behaviors | |
| CN107888388B (en) | Charging method and system for network acceleration service | |
| CN101567072A (en) | Method, device and system for connection of consultation service | |
| KR101648470B1 (en) | System for real time distributing information based on code and keyword by using smart terminal and method thereof | |
| WO2020234860A1 (en) | System and method for prediction of tv users engagement | |
| Rafter et al. | Inferring relevance feedback from server logs: A case study in online recruitment | |
| JP3719342B2 (en) | Hypertext analysis apparatus and method, and storage medium storing hypertext analysis program | |
| CN111651336B (en) | Data reporting method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160608 |
|
| RJ01 | Rejection of invention patent application after publication |