CN106469159A - A kind of big data analysis method based on viewing User Status, apparatus and system - Google Patents

A kind of big data analysis method based on viewing User Status, apparatus and system Download PDF

Info

Publication number
CN106469159A
CN106469159A CN201510507056.0A CN201510507056A CN106469159A CN 106469159 A CN106469159 A CN 106469159A CN 201510507056 A CN201510507056 A CN 201510507056A CN 106469159 A CN106469159 A CN 106469159A
Authority
CN
China
Prior art keywords
information
multimedia file
client terminal
played multimedia
label
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
Application number
CN201510507056.0A
Other languages
Chinese (zh)
Inventor
***
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
SHANGHAI MADHOUSE INFORMATION TECHNOLOGY Co Ltd
Original Assignee
SHANGHAI MADHOUSE INFORMATION TECHNOLOGY Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by SHANGHAI MADHOUSE INFORMATION TECHNOLOGY Co Ltd filed Critical SHANGHAI MADHOUSE INFORMATION TECHNOLOGY Co Ltd
Priority to CN201510507056.0A priority Critical patent/CN106469159A/en
Publication of CN106469159A publication Critical patent/CN106469159A/en
Pending legal-status Critical Current

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
    • G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
    • G06F16/9562—Bookmark management
    • 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 Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The present invention provides a kind of big data analysis method based on viewing User Status, and described analysis result is used for releasing news, and the described information that is published is corresponding with described analysis result, including：Information to be released is determined based on the information parameter of played multimedia file；Obtain the status information of the client terminal being associated with described played multimedia file；Determine the transmission time sending described information to be released based on described status information.Correspondingly analytical equipment and system are also provided.The mode that the present invention is combined based on mutual information I and positive feature selection approach, analysis on the basis of time geography position model or time model determines the described transmission time, the technical scheme being determined the information issuing time based on viewing User Status is provided, make user can receive the related information of multimedia file that is current to it or will watching in the very first time, or receive described information after terminating viewing, improve the work efficiency of information publisher, optimize Consumer's Experience.

Description

A kind of big data analysis method based on viewing User Status, apparatus and system
Technical field
A kind of the invention belongs to information issues analysis system, in particular it relates to method and corresponding analytical equipment and analysis system carrying out big data analysis based on viewing User Status.
Background technology
High speed development with wireless Internet technologies, number of users using the online of the mobile devices such as mobile phone, IPAD grows with each passing day, many enterprises, star start with wireless Internet to service for oneself, itself popularity is improved by wireless Internet, and then creation commercial value, and the first step that releasing advertisements are exactly enhanced the reputation is also most essential steps.
Existing advertiser is substantially and sends advertisement using fixed point input mode to mobile subscriber, for example, an advertisement position of the latter's mobile client had been bought in one advertiser and Tengxun's video cooperation, when user's one film of online viewing to terminate or midway suspends when, advertiser can issue star's vermicelli meet-and-greet advertisement on currently playing window, but this ad distribution mode is changeless to the advertisement that user delivers within a certain period of time, cannot be according to user's viewing state real-time change ad content, it is possibly even out-of-date advertisement, this does not have any effect for improving advertiser's popularity.On the other hand, advertiser is also possible to send advertisement with note form to the user's mobile device being currently in a certain region, all for advertiser advertisements are not often done screening ground and are disposably sent to all users by this ad distribution mode, cannot suitable advertisement customized to different user, have impact on efficiency and the precision of advertisement delivering.For example, show multi-section film in a certain cinema simultaneously, then advertiser to current at the cinema in all user's mobile devices have sent the multiple advertisements related respectively to multi-section film, but general user at the cinema in the same time only can watch a film, watch film closely related advertising message thus ignoring with this user if receive multiple movie related advertisement and be likely to cause read fatigue quickly.
Many times, advertiser needs a kind of method that viewing User Status can be optimized with analysis, reasonably can deliver different advertisements stage by stage according to the different conditions of viewing user.For example, after user have purchased a film ticket, advertiser can obtain the information of mobile terminal of this user based on cinemas Ticketing information, if this information of mobile terminal display user current just at the cinema around and current time begin also for some time from film, push the advertising messages such as merchant information, the entertainment information of cinema's periphery to user；If this information of mobile terminal display film prologue time, current time was at 10 points in the morning for 12 noon, push the diet advertisement information of cinema's periphery to user；If this information of mobile terminal display user pay special attention to a star in this film, before motion picture projection, the advertisement of vying to answer with lottery of this star relevant is then pushed to user, point out this advertisement again after motion picture projection, if user answers correctly, have an opportunity to participate in the vermicelli meet-and-greet of this star.For advertiser, the above-mentioned this method delivering advertisement is only the most efficient accurately advertisement delivering method, also allows for the puzzlement that user breaks away from magnanimity waste advertisements, only receives oneself advertising message interested.
At this stage, the extraordinary method of neither one solves the problems, such as mentioned above.In most cases, all advertisements of advertiser can only be not added with taxonomically being sent to customer mobile terminal, a kind of effectively analysis viewing User Status are not provided, precisely delivers the technical scheme of advertisement according to analysis result at times.
Content of the invention
Do not have in prior art to provide the technical problem analyzing the technical scheme determining the information issuing time based on viewing User Status to overcome, it is an object of the invention to provide a kind of big data analysis method based on viewing User Status, corresponding analytical equipment and analysis system.
According to an aspect of the present invention, provide a kind of big data analysis method based on viewing User Status, described analysis result is used for releasing news, described be published information corresponding with described analysis result it is characterised in that comprising the steps：
A. information to be released is determined based on the information parameter of played multimedia file；
B. obtain the status information of the client terminal being associated with described played multimedia file；
C. the transmission time sending described information to be released is determined based on described status information.
Preferably, multiple described information to be released form a set U, and described step a comprises the steps：
A1. using in multiple for described played multimedia file label informations with described played multimedia file dependency highest N number of primary label information as described information parameter, wherein N >=1；
A2. corresponding information to be released is determined in described set U based on described information parameter.
Preferably, described set U also includes the corresponding label information of described information to be released.
Preferably, described step a1 comprises the steps：
A11. calculate the degree of association I of the plurality of label information and a theme respectively, wherein said degree of association I passes through following mutual information formula and calculates acquisition：
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, and it is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) is the joint probability distribution of x and y；
A12. it is based on described degree of association I and obtain maximally related M label information, wherein M ＞ N with described theme；
A13. it is based on positive feature selection approach to select to determine that described N primary label information forms set U1 from described M label information.
Preferably, described forward direction feature selection approach is a greedy algorithm, and described step a13 comprises the steps：
- the degree of association I1 of described M label information and described set U1 is calculated respectively based on equation below：
Wherein, x is label information to be screened in described M label information；yjFor j-th label information in described set U1；
- selecting a minimum label of described I1 to be added in described set U1 for new label every time, the number of labels up in described set U1 reaches N number of.
Preferably, described step a comprises the steps：
A1 '. using lower for classification same in multiple for described played multimedia file label informations frequency of occurrence highest N number of primary label information as described information parameter, wherein N >=1；
A2. corresponding information to be released is determined in described set U based on described information parameter.
Preferably, described step a2 comprises the steps：
A21. one or more described information to be released are obtained from described set U based on described theme, described information to be released includes one or more label informations, and described label information is corresponding with described theme；
A22. it is based on natural language model, calculate the degree of association of one or more label informations that information parameter is included with described information to be released described in described set U1 by equation below：
Wherein, x is information parameter；X is label information.
A23. determine that the described information to be released of described degree of association highest M is preferably to release news, wherein M >=1.
Preferably, described step b comprises the steps：
B1. determine the client terminal being associated with described played multimedia file；
B2. obtain the status information of described client terminal, wherein said status information at least includes any one in following information or appoints multiple：
The geographical location information of-described client terminal；
The temporal information of-described client terminal；
The history information of-described client terminal.
Preferably, described step b also comprises the steps：
B3. obtain the status information of described played multimedia file, wherein said status information at least includes any one in following information or appoints multiple：
The reproduction time of-described played multimedia file；
The broadcasting place of-described played multimedia file.
Preferably, described step b1 comprises the steps：
B11. using the terminal playing described played multimedia file as described client terminal.
Preferably, described step b1 also comprises the steps：
B11 '. obtain the corresponding Internet registry information of described played multimedia file, multiple terminals that described Internet registry information is included are as described client terminal.
Preferably, described step c comprises the steps：
C1. the degree of association between the corresponding status information of described client terminal status information corresponding with described played multimedia file is calculated based on time geography position model, wherein said degree of association is calculated by equation below：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T is described played multimedia file broadcast start time；W is described client terminal current location；W is the broadcasting place of described played multimedia file；
C2. the push time pushing described information to be released to described client terminal is determined based on described degree of association F.
Preferably, described f (t, T) is based on equation below and calculates：
Wherein, described t and described T is based on 24 hours tabulating showing.
Correspondingly, described f (w, W) is based on equation below and obtains：
Wherein, k is gauged distance, and for indicating, described client terminal current location is corresponding with described played multimedia file to play the distance between place to | w-W |.
Preferably, described step c also comprises the steps：
C1 '. described client terminal current time and described played multimedia file broadcast start time are analyzed based on time model；
C2 '. the push time pushing described information to be released to described client terminal is determined based on analysis result.
According to a further aspect in the invention, also provide a kind of big data analytical equipment based on viewing User Status, described analysis result is used for releasing news, described be published information corresponding with described analysis result it is characterised in that include：
First determination device, it is used for determining information to be released based on the information parameter of played multimedia file；
First acquisition device, it is used for obtaining the status information of the client terminal being associated with described played multimedia file；
Second determination device, it is used for determining, based on described status information, the transmission time sending described information to be released.
Preferably, multiple described information to be released form a set U, and described first determination device includes：
First processing meanss, its be used for using in multiple for described played multimedia file label informations with described played multimedia file dependency highest N number of primary label information as described information parameter, wherein N >=1；
3rd determination device, it is used for determining corresponding information to be released in described set U based on described information parameter.
Preferably, described set U also includes the corresponding label information of described information to be released.
Preferably, described first processing meanss include：
Second processing device, it is used for calculating the degree of association I of the plurality of label information and a theme respectively, and wherein said degree of association I passes through following mutual information formula and calculates acquisition：
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, and it is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) is the joint probability distribution of x and y；
3rd processing meanss, it is used for obtaining maximally related M label information, wherein M ＞ N with described theme based on described degree of association I；
4th determination device, it is used for selecting to determine that described N number of primary label information forms set U1 from described M label information based on positive feature selection approach.
Preferably, described forward direction feature selection approach is a greedy algorithm, and described 4th determination device includes：
Fourth process device, it is used for calculating the degree of association I1 of described M label information and described set U1 respectively based on equation below：
Wherein, x is label information to be screened in described M label information；yjFor j-th label information in described set U1；
First choice device, it is used for each label selecting described I1 minimum and is added in described set U1 for new label, and the number of labels up in described set U1 reaches N number of.
Preferably, described first determination device also includes：
5th processing meanss, it is used for same classification time frequency of occurrence highest N number of primary label information in multiple for described played multimedia file label informations as described information parameter, wherein N >=1；
3rd determination device, it is used for determining corresponding information to be released in described set U based on described information parameter.
Preferably, described 3rd determination device includes：
6th processing meanss, it is used for obtaining one or more described information to be released from described set U based on described theme, and described information to be released includes one or more label informations, and described label information is corresponding with described theme；
7th processing meanss, it is used for based on natural language model, calculates the degree of association of one or more label informations that information parameter is included described in described set U1 with described information to be released by equation below：
Wherein, x is information parameter；X is label information.
5th determination device, it is used for determining that the described information to be released of described degree of association highest M is preferably to release news, wherein M >=1.
Preferably, described first acquisition device includes：
6th determination device, it is used for determining the client terminal being associated with described played multimedia file；
Second acquisition device, it is used for obtaining the status information of described client terminal, and wherein said status information at least includes any one in following information or appoints multiple：
The geographical location information of-described client terminal；
The temporal information of-described client terminal；
The history information of-described client terminal.
Preferably, described first acquisition device also includes：
3rd acquisition device, it is used for obtaining the status information of described played multimedia file, and wherein said status information at least includes any one in following information or appoints multiple：
The reproduction time of-described played multimedia file；
The broadcasting place of-described played multimedia file.
Preferably, described 6th determination device includes：
7th determination device, it is used for the terminal playing described played multimedia file as described client terminal.
Preferably, described 6th determination device also includes：
8th determination device, it is used for obtaining the corresponding Internet registry information of described played multimedia file, and multiple terminals that described Internet registry information is included are as described client terminal.
Preferably, described second determination device includes：
8th processing meanss, it is used for calculating degree of association between the corresponding status information of described client terminal status information corresponding with described played multimedia file based on time geography position model, and wherein said degree of association is calculated by equation below：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T is described played multimedia file broadcast start time；W is described client terminal current location；W is the broadcasting place of described played multimedia file；
9th determination device, it is used for determining, based on described degree of association F, the push time pushing described information to be released to described client terminal.
Preferably, described f (t, T) is based on equation below and calculates：
Wherein, described t and described T is based on 24 hours tabulating showing.
Correspondingly, described f (w, W) is based on equation below and obtains：
Wherein, k is gauged distance, and for indicating, described client terminal current location is corresponding with described played multimedia file to play the distance between place to | w-W |.
Preferably, described second determination device also includes：
First analytical equipment, it is used for analyzing described client terminal current time and described played multimedia file broadcast start time based on time model；
Tenth determination device, it is used for determining, based on analysis result, the push time pushing described information to be released to described client terminal.
According to another aspect of the invention, also provide a kind of big data analysis system based on viewing User Status, described analysis result is used for releasing news, described be published information corresponding with described analysis result it is characterised in that include：
First server, it is used for determining information to be released based on the information parameter of played multimedia file, obtains the status information of client terminal being associated with described played multimedia file and determines the transmission time sending described information to be released based on described status information；
Second server, it is used for determining the client terminal being associated with described played multimedia file；
Described analysis system also includes the analytical equipment according to foregoing invention content any one of second aspect, and wherein, described first server, second server and described analytical equipment communicate.
Brief description
By reading detailed description non-limiting example made with reference to the following drawings, the other features, objects and advantages of the present invention will become more apparent upon：
Fig. 1 illustrates according to the first embodiment of the present invention, the flow chart of the big data analysis method based on viewing User Status；
Fig. 2 illustrates according to the second embodiment of the present invention, the flow chart of the big data analysis method based on viewing User Status；
According to the third embodiment of the invention Fig. 3 illustrates, the flow chart of the big data analysis method based on viewing User Status；
Fig. 4 illustrates according to the fourth embodiment of the invention, the structure chart of the big data analytical equipment based on viewing User Status；And
Fig. 5 illustrates the network topology structure schematic diagram of typical application scenarios according to the present invention.
Specific embodiment
In order to preferably make technical scheme clearly show, the invention will be further described below in conjunction with the accompanying drawings.
Skilled artisan understands that, issue, in order to solve information, the technical problem that analysis system targetedly cannot release news according to user's viewing state, technical program of the present invention lies in providing a kind of solution that can precisely issue to point content the information closely related with user's viewing content according to the viewing state of different user at times.The technical scheme being provided by the present invention, can optimize the information quality of reception of terminal use significantly, be greatly enhanced office efficiency and the precision of information publisher simultaneously.Specifically, in a preferred embodiment of the invention, preferably release news by degree of association I determination is corresponding with being played multimedia file, and determined to the client terminal described Best Times preferably releasing news of transmission based on analysis methods such as time geographic model or time models.
Fig. 1 illustrates that described analysis result is used for releasing news according to the first embodiment of the present invention, the flow chart of the big data analysis method based on viewing User Status, and described to be published information corresponding with described analysis result.Specifically, in the present embodiment, step S101 is first carried out, calculates the degree of association I of the played multiple label information of multimedia file and a theme respectively.More specifically, described label information includes the characteristic information of described played multimedia file, at least includes performers and clerks' information of described played multimedia file, brand sponsored messages, prevailing scenario information etc..More specifically, described label information can be by scanning acquisition to described played multimedia file frame by frame.Further, described theme includes the feature general introduction in a certain respect of described played multimedia file, corresponding with described label information.Preferably, described degree of association is used for weighing the correlation degree of multiple described label informations theme corresponding with described played multimedia file, and described label information is higher with described degree of subject relativity, then association between the two is tightr.
In a preference, described degree of association I passes through following mutual information formula and calculates acquisition：
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, and it is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) is the joint probability distribution of x and y.
Preferably, described benchmark label information y is corresponding with described theme, usually can reflect the key word of described theme feature.Preferably, described benchmark label information y can also directly be represented with described theme.Preferably, described mutual information is used for indicating estimating between described label information and described theme, if described label information belongs to described theme corresponding types, its mutual information result of calculation is maximum.If it will be appreciated by those skilled in the art that label information frequency of occurrences in a theme high and in other themes the frequency of occurrences low then it is assumed that described label information mutual information corresponding with described theme is larger.Preferably, described KL function is used for characterizing the relation between two distributions.Preferably, if I is [x, y]=0, then represent that described label information to be screened is unrelated with described theme, understand from information theory view, now described big data analysis system is observed the quantity of information of described label information to be screened in the case of observing described theme and is equal to and is directly observed the quantity of information of described label information to be screened, for example, I [action, animation]=0 represents that the frequency that actions tab is observed in " animation " class multimedia file is 0.Skilled artisan understands that, described big data analysis system weighs the correlation degree of the multiple label of described played multimedia file and described theme based on described degree of association I, and the plurality of label and corresponding degree of association I are stored in a storage device for subsequently analyzing further.Preferably, described storage device at least includes any one in flash memory, background thread, Flash, cloud storage device etc. or appoints multiple.
Subsequently into the execution of step S102, the maximally related M label information with described theme is obtained based on described degree of association I.Specifically, described big data analysis system is filtered out and described degree of subject relativity highest M label information, wherein M ＞ N based on described mutual information I [x, y] from the multiple information labels of described multimedia file.More specifically, described label information is stored in a storage device with corresponding degree of association I.Further, described label information can correspond to multiple described degree of association I, each described degree of association is corresponding with a theme of described played multimedia file, for example, described played multimedia file includes multiple themes " risk " simultaneously, " action ", " magical ", then described big data analysis system calculates the degree of association I between label information " chocolate " and the plurality of theme respectively, and described result of calculation is stored in described storage device, when described big data analysis system needs to analyze the relatedness between described some particular topic of played multimedia file and the plurality of label information, then described degree of association I numerical value and corresponding label information accordingly are transferred from described storage device based on described theme.Preferably, described N is described played multimedia file corresponding informance number of parameters, and it is used for indicating in the multiple label information of described played multimedia file N number of primary label information, wherein N >=1 with described played multimedia file dependency highest.Preferably, described big data analysis system is based on the described information parameter determination to be released information corresponding with described played multimedia file.Preferably, described information to be released is sent to one or more client terminals by an information delivery server.In a preference, described big data analysis system is based on described degree of association I, and to described played multimedia file, all label informations arrange, preferably using M maximum for degree of association I numerical value described under same subject label information as described label information, thus preliminary screening is carried out to all label informations of described played multimedia file.Preferably, described M=20, those skilled in the art can also become according to actual needs and dissolve more embodiments, and this has no effect on the technology contents of the present invention.Preferably, described M label information and corresponding degree of association I numerical value are stored in described storage device.
Next enter step S103, N number of primary label information composition set U1 is selected as information parameter, wherein M ＞ N from described M label information based on positive feature selection approach.Specifically, described M label information is stored in described storage device.More specifically, described forward direction feature selection approach is a greedy algorithm, and it is used for carrying out postsearch screening to described M label information, thus excluding excessively approximate useless label in described M label information.Further, described set U1 includes N number of information parameter with described played multimedia file degree of subject relativity highest, and it preferably chooses the minimum N number of label information composition of dependency each other from described M label information.Skilled artisan understands that, may lead under described played multimedia file same subject, obtain the identical or approximate label information of multiple and described degree of subject relativity I numerical value based on the technical scheme that described degree of association I obtains described label information in above-mentioned steps S101 and step S102 simultaneously, if only using M maximum label information of described degree of association I numerical value as described information parameter, be unfavorable for that described big data analysis system is weighed to described played multimedia file comprehensively, for example, described played multimedia file is based on the maximum label information of " love " degree of subject relativity I and is respectively：" jewelry ", " gem ", " jade ", " platinum ", " diamond ring ", degree of approximation highest wherein between " jewelry ", " gem " and " jade ", if directly using above-mentioned label information as described information parameter, described label information " jewelry ", " gem " and " jade " is likely to corresponding same information to be released, have impact on the diversified demand of described information parameter, be unfavorable for that described information publisher server is based on limited described information gain of parameter described information to be released as much as possible.In a preference，In order to exclude excessively approximate label information in described M label information，Described big data analysis system introduces described forward direction feature selection approach on the basis of described degree of association I and carries out postsearch screening to described M label information，Described big data analysis system is extracted any one of described M label information label information first and is put into described set U1，Then described M-1 label information is carried out degree of association I1 calculating with the label information in described set U1 respectively，The label information that a label information minimum with label information correlation I1 in described set U1 is new is selected to be added in described set U1 every time，The like，Until label information quantity reaches predetermined value N described in described set U1，The information parameter that then now the N number of described label information in described set U1 needs for described information publisher server，Can be used for determining described information to be released.Preferably, described big data analysis system calculates the degree of association I1 of described M label information and each label information in described set U1 respectively based on equation below：
Wherein, x is label information to be screened in described M label information；yjFor j-th label information in described set U1, N is label information content in described set U1.Preferably, described I1 function is based on natural language model and calculates described x and described yjDegree of association.Preferably, described 11 functions are also based on mutual information formula described in above-mentioned steps S101 and calculate acquisition.Preferably, described big data analysis system selects a minimum label information of described I1 to be added in described set U1 for new label information every time, and the number of labels up in described set U1 reaches N number of.It will be appreciated by those skilled in the art that N number of label information includes two characteristics described in described set U1, on the one hand described N number of label information degree of association I highest corresponding with described played multimedia file theme；On the other hand described N number of label information degree of association I1 each other is minimum，Described big data analysis system obtains described set U1 by the method that described mutual information formula and described forward direction feature selection combine，N number of information parameter described in described set U1 is made broadly to be related to the every aspect of public's daily life under described played multimedia file theme as far as possible，So that described information publisher server can obtain as multiple as possible described information to be released based on described N number of information parameter，And the difference between described information to be released maximizes，This is the technical scheme not adopted in prior art，Permission information publisher can obtain information to be released based on described played multimedia file in different field，Skilled artisan understands that，The more probabilities that it is noticed by a certain specific user during formula of casting net is issued in field that described information to be released is related to are bigger，Thus extending the potential audience scope of described information to be released，Improve the work efficiency of described information publisher，Also optimize the operating experience of user，Improve user and receive the probability with oneself hobby relevant information.Preferably, described N=10, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.Preferably, described set U1 and corresponding theme are stored in described storage device.
Then execution step S104, based on natural language model, calculates the degree of association of described information parameter and multiple label informations in set U.Specifically, multiple described information to be released form described set U, are stored in described storage device.More specifically, described set U also includes described label information, corresponding with described information to be released.Further, described label information includes described information principal character to be released, and for example, described information to be released corresponds to brand name, the corresponding mouthpiece's title of described information to be released or the corresponding descriptor of described information to be released etc..Preferably, described information to be released may correspond to one or more described label informations, and for example, in described set U, an information to be released includes " light ", " risk " and " Huang Xiaoming " three label informations simultaneously.Preferably, described big data analysis system weighs degree of association between described information parameter and multiple described label information respectively based on described natural language model, thus obtaining the relevancy ranking of multiple described information to be released and described information parameter in described set U.Preferably, described information parameter obtains from described set U1, described big data analysis system extract from described set U1 successively an information parameter calculate its with described set U in multiple described label informations degree of association, finally obtain the degree of association of information to be released described in all informations parameter and described set U in described set U1.In a preference, described big data analysis system is primarily based on described played multimedia file theme and obtains one or more described information to be released from described set U, is then based on equation below and calculates described degree of association：
Wherein, x is information parameter, and X is label information, and N includes number of elements for described set U1, n includes number of elements for described set U, and f function is based on natural language model and calculates acquisition, for example, f (Huang Xiaoming, that English)=0, f (love, emotion)=0.5.Skilled artisan understands that，Described set U1 includes N number of described information parameter，Correspondingly，Described set U includes multiple described information to be released，And described information to be released also includes one or more label informations，Then described big data analysis system needs corresponding with described information to be released for described N number of information parameter multiple label information Integrated comparative，So that the result of calculation of described degree of association is more accurate，For example，Described played multimedia file includes a theme " performers and clerks "，Calculate the described set U1 of acquisition by above-mentioned steps S101 to above-mentioned steps S103 and include described information parameter " A "、" B " and " C "，And include information a corresponding label information " A " to be released in described set U，One information b corresponding label information " A " and " B " to be released，Then described information a to be released corresponds to degree of association F=f (A，A)+f(B，A)+f(C，A)=1+0+0=1，Described information b to be released corresponds to degree of association F=f (A，A)+f(A，B)+f(A，C)+f(B，A)+f(B，B)+f(B，C)=1+0+0+0+1+0=2，Then described information b to be released and described set U1 information parameter degree of association are more than described information a to be released，Described big data analysis system can be made more accurate degree of association based on technique scheme to described information parameter and described information to be released and judge.Preferably, described relatedness computation result is stored in described storage device, corresponding with described information to be released.
Next enter the execution of step S105, determine that the corresponding information to be released of described degree of association highest M label information is preferably to release news, wherein M >=1.Specifically, described correlation score is stored in described storage device, corresponding with described information to be released.More specifically, described big data analysis system preferably will be released news as described with described set U1 degree of association highest M information to be released in described set U.Further, described preferably release news including coding information and described label information, described coding information includes the information write based on Javascript language, it is used for described preferably releasing news is converted into one section of code and is stored in described storage device, described coding information is decoded as after word, image or both combinations to be sent to client terminal by described information publisher server, thus preferably releasing news displaying in front of the user by described.Preferably, described M=5, those skilled in the art can also become according to actual needs and dissolve more embodiments, and this has no effect on the technology contents of the present invention.Preferably, described preferably releasing news is stored in described storage device with code form.Preferably, described preferably releasing news can be stored in described storage device in the form of word, image or both combinations, then described information publisher server can directly transmit without decoding operate when preferably releasing news to the transmission of described client terminal is described.
Subsequently into the execution of step S106, obtain the corresponding Internet registry information of described played multimedia file, multiple terminals that described Internet registry information is included are as described client terminal.Specifically, described Internet registry information includes the corresponding history information of described client terminal and described client terminal corresponding user information.More specifically, described user profile includes identity information, preference information of filling in when user is registered on an internet site etc., the name of such as user, contact method, age, hobby, viewing type etc. based on described client terminal.Further, described history information includes the internet hunt record in described client terminal a period of time recently and browses record.Preferably, described internet site is an internet electronic ticket-booking system, communicates with cinemas box office background server.Preferably, described Internet registry information also includes Transaction Information on described internet site for the nearest n times of described client terminal.In a preference, described Transaction Information includes the Ticketing information that described client terminal is produced based on register account number, described register account number is corresponding with described internet electronic ticket-booking system, described Ticketing information includes described played multimedia file and corresponds to title, reproduction time, play place etc., for example, user is after described internet electronic ticket-booking system completes registration, 2 film tickets by described systems buying, the multimedia file that described film ticket corresponding is being play or will play, then described big data analysis system can obtain the Internet registry information of described user based on Ticketing information described in user, reproduction time including the corresponding played multimedia file of this booking of described user, play the fileinfos such as place and the corresponding client terminal information of user's booking etc..Preferably, described big data analysis system periodically obtains described Internet registry information from described cinemas box office background server, and the described cycle is 1 week/time, and those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.Preferably, described big data analysis system can also be communicated with described cinemas box office background server in real time, thus obtaining described Internet registry information in time.In another preference, described internet site can also be online playing platform, it is used for providing, to the corresponding user of described client terminal, resource and the window watching described played multimedia file online, described online playing platform includes charge or free-form, described user can browse described played multimedia file in described online playing platform after completing to be registered as member on the platform, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.Preferably, described N=1, those skilled in the art can become according to actual needs and dissolve more embodiments, and this has no effect on the technology contents of the present invention.
Next execution step S107, obtains described client terminal and the status information of described played multimedia file respectively.Specifically, described status information at least includes any one in the history information of geographical location information, the temporal information of described client terminal or described client terminal of described client terminal or appoints multiple.More specifically, the status information of described played multimedia file includes any one in broadcasting place of the reproduction time of described played multimedia file, described played multimedia file etc. or appoints multiple.Further, described status information is stored in described storage device, corresponding with described client terminal or described played multimedia file title.Preferably, described geographical location information is based on GPS (global positioning system) and obtains, corresponding with the gps signal that the GPS device that described client terminal includes sends in real time.Preferably, described temporal information is based on described client terminal present system time and obtains.Preferably, described temporal information is also based on the acquisition of corresponding Greenwich Mean Time of described client terminal current geographic position.Preferably, described history information is browsed based on described client terminal historical internet or searches for record and obtains.In a preference, described information publisher server is using the terminal playing described played multimedia file as after described client terminal, send solicited message to described client terminal, client terminal status information described in acquisition request, described client terminal receives after described solicited message to described information publisher server send feedback information, described feedback information is corresponding with described solicited message, including described client terminal status information.In another preference, after described client terminal receives described solicited message, prompt the user with described solicited message, and select to send described feedback information to described information publisher server based on user, if user selects to agree to, described feedback information includes thinking corresponding described client terminal status information with solicited message；If user selects refusal, described feedback information item is sky, and described information publisher server cannot obtain described client terminal corresponding statess information based on described feedback information.Preferably, the status information based on multimedia file played described in the file bag content obtaining of described multimedia file for the described information publisher server.Preferably, if the terminal of currently playing described played multimedia file is described client terminal, the corresponding geographical location information of described playback terminal play place corresponding with described played multimedia file is same content actually.Preferably, if described client terminal does not open described GPS location function, then described client terminal is after receiving described solicited message, prompt the user with the Geoposition Request information that described solicited message includes, and user input result is fed back to described information publisher server as described feedback information.
Next enter step S108, the degree of association F between described client terminal corresponding statess information and described played multimedia file corresponding statess information is calculated based on time geographic model.Specifically, described status information is stored in described storage device, corresponding with described client terminal or described played multimedia file.More specifically, described degree of association F is used for weighing the push described push time preferably releasing news.Preferably, the corresponding result of calculation of described degree of association F is stored in described storage device.Preferably, described degree of association F is based on equation below and calculates：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T is described played multimedia file broadcast start time；W is described client terminal current location；W is the broadcasting place of described played multimedia file.In a preference, described f function is used for indicating the degree of correlation between two variables, described f (t, T) is used for indicating the distance between described client terminal current time and described played multimedia file broadcast start time, is calculated based on equation below and obtains：
Wherein, described t and described T is based on 24 hours tabulating showing.Described t from described T more away from then described f numerical value bigger.For example, f (14,16)=0.17, f (15,15)=0.Preferably, described f (t, T) numerical value is between 0～1.
Preferably, described f (w, W) is used for indicating that described client terminal current location is corresponding with described played multimedia file plays the distance between place, is calculated based on equation below and obtains：
Wherein, k is gauged distance, and for indicating, described client terminal current location is corresponding with described played multimedia file to play the distance between place to | w-W |.Preferably, the more near then described f (w of described distance, W) numerical value is less, described f (w, W) value also between 0～1, preferably, described K=10000 rice, those skilled in the art can also become according to actual needs and dissolve more embodiments, to guarantee described f (w, W) numerical value between 0～1.Preferably, the span of described degree of association F is also between 0～1.In another preference, described t and described T is based on tabulation in 12 hours and shows, then described f (t, T) function is accordingly being calculated after described t and described T is converted to first the time scale under 24 hours systems when calculating.
Last execution step S109, determines the push time pushing described information to be released to described client terminal based on described degree of association F.Specifically, described degree of association F result of calculation is stored in described storage device, corresponding with described client terminal corresponding statess information and described played multimedia file corresponding statess information.More specifically, the numerical value of described degree of association F is between 0～1.Preferably, described degree of association F is used for representing time and the distance relation between described client terminal and described played multimedia file, if described degree of association F numerical value bigger then it represents that described client terminal is apart from described played multimedia file reproduction time, time point occurs in described played multimedia file and plays near place earlier；Correspondingly, if described degree of association F numerical value is less, represent that described client terminal reaches described played multimedia file and plays place when closing on described played multimedia file reproduction time.In a preference, described big data analysis system presets first threshold θ to determine the described push time, if F is ＞ θ, then judge that the corresponding user of described client terminal occurs in described played multimedia file in the more early time and plays near place, using described client terminal current time as the described push time, and described information publisher server is called to send described information to be released to described client terminal；If F is ＜ θ, then judge that the corresponding user of described client terminal closes on and just occur near described played multimedia file broadcasting place when described played multimedia file commences play out, described played multimedia file is terminated reproduction time as the described push time, after user terminates to watch described played multimedia file, send described information to be released to described client terminal again.Skilled artisan understands that, compared with prior art, technical scheme described in the present embodiment allows described big data analysis system to release news to user based on user and described played multimedia file state, make push time that described big data analysis system releases news to user and to push content more intelligent, can be different user customization personalized ground pushed information, for example, described played multimedia file is a cinema projection film, if the corresponding user of described client terminal more early time point before film prologue occurs near movie theatre, then judge to push the pushed information related to described film in current time to described user；If described user has just arrived movie theatre before film prologue, judge that terminating rear line in film projection pushes the pushed information related to described film.Preferably, described pushed information includes described played multimedia file schedule information, played multimedia file interactive information, the embedded brand message of played multimedia file etc. so that user can have a understanding by described pushed information to described played multimedia file；On the other hand, described pushed information can also be LBS advertising message, it is used for pushing the advertising messages such as a range of food and drink of presently described client terminal geographical position periphery, amusement to described client terminal, thus guiding the corresponding user of described client terminal can have a rational leisure place to go within this period before waiting described played multimedia file to start to show, drive periphery level of consumption simultaneously, improve the work efficiency of information publisher, optimize the operating experience of user.Preferably, described information publisher server is also based on the described push time to the different pushed information of described client terminal transmission, for example, pushes film schedule information before film projection, film is acted the leading role and represented advertising message etc.；And push with lottery vie to answer information related to film etc. after film projection, skilled artisan understands that, step described in the present embodiment is the technical scheme that prior art does not adopt, make the information that user receives more targetedly, the time of the described information of user's reception simultaneously is also more reasonable, improve the success rate that information is paid close attention to by user, drastically increase the work efficiency of information publisher.Preferably, described θ=0.5, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.
In a change case of the present embodiment, degree of association described in described step S104 is compared based on described natural language model, correspondingly, described information parameter first passes through normalized in advance so that consistent with label information in described set U to the statement of a certain class word in described information parameter.Preferably, described normalized is processed to described information parameter based on a vocabulary contrast indication storehouse, described vocabulary contrast indication storehouse includes one or more key words, each described key word corresponds to multiple similar word, for example, key word " jewelry " can correspond to multiple similar word " gem ", " jade ", " necklace ", " diamond " etc..Skilled artisan understands that，Described information parameter is during being extracted from described played multimedia file based on artificially defined or computer-assisted analyses，Produce nuance because of the language convention of different geographical or different nationalities，Same vocabulary there may be substantial amounts of similar word or approximate word，On the other hand，Described played multimedia file may be split simultaneously processed to improve efficiency to multiple big data Analysis servers by described big data analysis system，Then different server processes the described information parameter obtaining possible mutual common factor but is not quite similar，Then described information parameter is compared by described big data analysis system with described vocabulary contrast indication storehouse，If there being multiple described information parameters can correspond to same key word based on described vocabulary contrast indication storehouse，Then the plurality of information parameter replaced in described key word，Thus simplifying to described information parameter sets，Make N number of information parameter described in described set U1 farthest unconnected from each other while embodying described played multimedia file theme.Correspondingly, information corresponding label information to be released described in described set U is extracted also based in the key word of described vocabulary contrast indication storehouse, then described information parameter only need to be carried out homogeny with the label information in described set U and compares the degree of association being achieved with described information parameter and described label information by described big data analysis system, if described information parameter is identical with described label information, judge the to be released information positive correlation corresponding with described label information of described information parameter, now described degree of association is 1；If described information parameter is differed with described label information, judge that described information parameter to be released information corresponding with described label information is negatively correlated, now degree of association is -1, one may finally be obtained by 1 and -1 set forming, described set is used for indicating the dependency of described information parameter and described information to be released after described information parameter is compared with multiple described label informations in described set U.In another preference, described vocabulary contrast indication storehouse is preset in one or more described big data Analysis servers, described big data analysis system extracts the plurality of label information based on described vocabulary contrast indication storehouse from described played multimedia file, thus eliminating the additional step that described big data analysis system needs the plurality of label information is normalized.Skilled artisan understands that, compared with technical scheme described in above-described embodiment, technical scheme described in this change case is by pre-setting unified standard to described label information, different personnel or different big data Analysis server are analyzed to described played multimedia file based on same standard, improve big data analysis efficiency.
In another change case of the present embodiment, preferably release news described in described step S105 and can also be stored in described storage device with internet link address format, described internet link address is corresponding with a webpage or website, described webpage includes the described actual content preferably releasing news such as word, image, video, and described website is made up of one or more described webpages.Preferably, described internet link address is sent to after described client terminal by described information publisher server, by clicking on described internet link address, user indicates that the currently displaying window of described client terminal jumps to the described actual content page that preferably releases news, thus browsing to described preferably releasing news.In another preference, if described client terminal current window corresponds to program not support of page turn, or user carries out page jump in currently displaying window after being not intended to click on described internet link address, then described client terminal is based on after user clicks on described internet link address and can not be redirected in currently displaying window, but call corresponding processing routine to realize page jump by background thread, when user wish to browse described preferably release news the page when only need to transfer corresponding program from background thread.Preferably, described preferably releasing news is sent to described client terminal by short message way, and described short message is based on mobile data flow and sends, such as note, multimedia message etc..Preferably，Described short message is also based on wireless Internet and sends，Such as wechat、Microblogging etc.，Skilled artisan understands that，Greatly developing with wireless Internet technologies，Described client terminal corresponding user gradually become accustomed to receive and send messages by described wireless Internet，Then described information publisher server can send described short message based on wireless Internet to the client terminal being provided with correspondingly received device，Wherein said reception device is an internet communication platform，It is pre-installed in described client terminal background program，For example，If wechat is provided with advance on described client terminal，Then described information publisher server preferably sends in described wechat and described preferably releases news，So that described information published method is accustomed to closer to user's daily life，Improve user to the described attention rate being published information，The information simultaneously reducing described information publisher to a certain extent issues cost.
In another change case of the present embodiment, described step S103 could alternatively be " extracting primary label information composition set U1 from multiple label information set respectively based on positive feature selection approach ".Specifically, described label information set includes M label information and the corresponding played multimedia file theme of described label information.More specifically, described big data analysis system is based on the multiple theme of described played multimedia file and obtains multiple label information set, and each described label information set corresponds to a theme of described played multimedia file respectively.Skilled artisan understands that, described played multimedia file potentially includes multiple themes, for example described played multimedia file includes theme " risk " simultaneously, " performer ", " brand " etc., obtain M label information maximally related with described theme after the calculating based on described degree of association I for each described theme and form described label information set, then described big data analysis system obtains N number of primary label information respectively from each described label information set, N=L1+L2+ ...+Ln, wherein L1 is L primary label information in the 1st label information set, corresponding with theme 1；Ln is L primary label information in n-th label information set, corresponding with theme n.Preferably, L1=L2=...=Ln, then described big data analysis system extract from each described label information set equal number label information form described set U1.Preferably, described big data analysis system is also based on the label information described set U1 of composition that normal distyribution function extracts varying number from each described label information set, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.
In another change case of the present embodiment, described step S106 could alternatively be " using the terminal playing described played multimedia file as client terminal." specifically, described client terminal is used for receiving and described preferably releases news.More specifically, described client terminal includes any one in mobile phone, IPAD, notebook computer, TV etc. or appoints multiple.Further，Described client terminal also includes and plays the display terminal that described played multimedia file terminal is connected，Described display terminal includes LED display、LCDs etc.，Communicated based on transferring device with described playback terminal，Described transferring device includes TV signal R-T unit、In short distance communication device etc. any one or appoint multiple，User by wired or wireless way, described playback terminal is connected with described display terminal thus playing described played multimedia file on described display terminal，Then user can select to receive on described display terminal and described preferably release news，Skilled artisan understands that，Prior art can realize the wirelessly or non-wirelessly communication between family multi-media equipment，After multiple described multimedia equipments are interconnected to form a communication network based on described wired or wireless mechanics of communication，On other arbitrarily individual multimedia equipment counterpart terminals that the multimedia file being currently played on one playback terminal can be shown in described communication network by user，Then described information publisher server can be using all multimedia equipment counterpart terminals in described communication network all as described client terminal，User can receive on all multimedia equipment counterpart terminals in described communication network and described preferably release news，Preferably can also release news based on oneself hobby selection one or more of multimedia equipment counterpart terminals reception is described.In this change case, because described client terminal is the terminal of currently playing described played multimedia file, correspondingly, between client terminal current location described in step S108 described in above-described embodiment and described played multimedia file corresponding broadcasting place, distance is 0, then described f (w, W)=1；And described client terminal current time is identical with described played multimedia file broadcast start time; then described f (t; T)=0; correspondingly; degree of association F=0 between the information of client terminal corresponding statess described in step S108 described in above-described embodiment and described played multimedia file corresponding statess information, then described information publisher server preferably suspend or send to described client terminal after terminating to play described played multimedia file in user and described preferably release news.
Fig. 2 illustrates according to the second embodiment of the present invention, the flow chart of the big data analysis method based on viewing User Status, and described analysis result is used for releasing news, and described to be published information corresponding with described analysis result.Specifically, in the present embodiment, step S111 is first carried out, using in multiple for played multimedia file label informations with described played multimedia file topic relativity highest N number of primary label information as information parameter.More specifically, described label information and described played multimedia file dependency are based on degree of association I and weigh, and described degree of association I is calculated by equation below：
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, and it is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) is the joint probability distribution of x and y.Preferably, the multiple label information of described played multimedia file is based on described degree of association I sequence, and described big data analysis system carries out postsearch screening, wherein M ＞ N to degree of association I highest M label information described in the plurality of label.Preferably, described postsearch screening includes excluding excessive useless label information in described M label information based on positive feature selection approach, using described the selection result as described information parameter.Further, those skilled in the art may be referred to step S101 described in above-mentioned embodiment illustrated in fig. 1, described step S102 and described step S103, will not be described here.
Subsequently into the execution of step S112, corresponding multiple label information is searched in described set U based on described information parameter.Specifically, described set U includes one or more information to be released and the corresponding label information of described information to be released.More specifically, described label information includes the principal character information of described information to be released, and for example, described information to be released corresponds to brand name, the corresponding mouthpiece's title of described information to be released or the corresponding descriptor of described information to be released etc..Further, each described information to be released may correspond to one or more described label informations.Preferably, described information parameter includes N number of label information and described set U1 correspondence theme in described set U1.Preferably, label information in described set U1 is compared by described big data analysis system successively with the label information in described set U, if have in multiple label informations that in described set U, an information to be released includes one identical or approximate with a label information in described set U1, described information to be released is released news as candidate.In a preference, by the information parameter in described set U1, all information corresponding label information to be released are compared described big data analysis system with described set U, and the candidate of acquisition are released news and corresponding label information Store is in a storage device.In another preference, described set U is equally based on described subject classification, each described theme includes multiple described information to be released and the corresponding one or more label informations of described information to be released, then described big data analysis system compares, based on described played multimedia file theme, multiple label informations that multiple accordingly described information to be released include in described set U, thus obtain one or more described candidates releasing news.
Next execution step S113, calculates the degree of association of described information parameter and described label information based on natural language model.Specifically, described label information includes described information principal character to be released, and for example, described information to be released corresponds to brand name, the corresponding mouthpiece's title of described information to be released or the corresponding descriptor of described information to be released etc..More specifically, described degree of association is based on equation below and calculates：
Wherein, x is information parameter, and X is label information, and N includes number of elements for described set U1, and n includes number of elements for described set U, and f function is based on natural language model and calculates acquisition.Further, those skilled in the art may be referred to step S104 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Then execution step S114, determines that the corresponding information to be released of described degree of association highest M label information is preferably to release news, wherein M >=1.Specifically, described correlation score is stored in described storage device, corresponding with described information to be released.More specifically, described big data analysis system preferably will be released news as described with described set U1 degree of association highest M information to be released in described set U.Further, those skilled in the art may be referred to step S105 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Next execution step S115, using the terminal playing described played multimedia file as client terminal.Specifically, described client terminal is used for receiving and described preferably releases news.More specifically, described client terminal includes any one in mobile phone, IPAD, notebook computer, TV etc. or appoints multiple.Further, described client terminal also includes and plays the display terminal that described played multimedia file terminal is connected, described display terminal includes LED display, LCDs etc., communicated based on transferring device with described playback terminal, described transferring device includes TV signal R-T unit, in short distance communication device etc. any one or appoint multiple, described playback terminal is connected with described display terminal thus playing described played multimedia file on described display terminal based on wired or wireless way by user by described transferring device, then user can select to receive on described display terminal and described preferably release news.In a preference, user selects described LCDs as described client terminal, described LCDs are connected based on HDMI high-definition data line with a transferring device, described transferring device is connected based on WLAN with described playback terminal, then user can browse described played multimedia file in described LCDs.In another preference, described transferring device is pre-installed in described display terminal, and described playback terminal passes through wired or wireless way directly and described display terminal is to communication, carries out mutually data exchange and file transmission.
Subsequently into the execution of step S116, obtain described client terminal and the status information of described played multimedia file respectively.Specifically, described status information at least includes any one in the history information of geographical location information, the temporal information of described client terminal or described client terminal of described client terminal or appoints multiple.More specifically, the status information of described played multimedia file includes any one in broadcasting place of the reproduction time of described played multimedia file, described played multimedia file etc. or appoints multiple.Further, those skilled in the art may be referred to step S107 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Next enter the execution of step S117, described client terminal current time and described played multimedia file broadcast start time are analyzed based on time model.Specifically, described client terminal current time is based on described client terminal background operating system and obtains.More specifically, described played multimedia file broadcast start time is based on described played multimedia paper document information acquisition.Further, described played multimedia paper document information also includes described played multimedia file and terminates reproduction time, and described end reproduction time can be obtained based on described played multimedia file broadcast start time and the total playing duration of described played multimedia file.Preferably, described played multimedia file broadcast start time and described client terminal current time are using with standard processed for the moment.Preferably, when described, standard is 24 hours systems, if standard recording when described played multimedia file broadcast start time adopts different from described client terminal current time, the described big data analysis system described time first is converted to the time representing under system 24 hours and then is analyzed.Preferably, described time model is analyzed based on equation below：
Wherein, Tt is the push time；T is described played multimedia file broadcast start time；T is described client terminal current time；β is constant.In a preference, described β=0, then described big data analysis system judges the precedence relationship between described played multimedia file broadcast start time and described client terminal current time, if it is t-T ＞ 0 that described played multimedia file broadcast start time is later than described client terminal current time, judge in A time point to described client terminal pushed information；If described played multimedia file broadcast start time early than or be t-T≤0 equal to described client terminal current time, then judge in B time point to described client terminal pushed information, for example, described played multimedia file commenced play out due to 14 points of today and the corresponding background operating system current time of described client terminal is at 10 points, then judge in A time point to described client terminal pushed information.It will be appreciated by those skilled in the art that described β can also become according to actual needs dissolves more embodiments, this has no effect on the technology contents of the present invention.Preferably, described it is stored in described storage device based on the analysis result of time model.In another preference, described analysis result preferably included described B time point as the described push time, skilled artisan understands that, step S115 described in the present embodiment is using the terminal of currently playing described played multimedia file as described client terminal, then described big data analysis system start to analyze described played multimedia file and described client terminal status information and temporal information when being analyzed described played multimedia file have begun to play, then described big data analysis system preferably using described B time point as the described push time.
Last execution step S118, determines the push time pushing described information to be released to described client terminal based on analysis result.Specifically, in described storage device, described storage device at least includes any one in flash memory, background thread, Flash, cloud storage device etc. or appoints multiple described information Store to be released.More specifically, described analysis result also is stored in described storage device.Preferably, the described push time is based on time point A and described B described in above-mentioned steps S117 and determines.In a preference, if described played multimedia file broadcast start time is later than described client terminal current time, judge in current time to described client terminal pushed information；If described played multimedia file broadcast start time early than or be equal to described client terminal current time; judge described played multimedia file terminate play after again to described client terminal pushed information.In another preference, described big data analysis system preferably pushes described information to be released when described played multimedia file suspends or terminates to play to described client terminal, for example, user opens after a multimedia file browsed based on a terminal, described big data analysis system is using the terminal of currently playing described played multimedia file as described client terminal, now described played multimedia file broadcast start time is already earlier than described client terminal current time, then described big data analysis system judges in B time point to described client terminal pushed information, described information to be released can be received when user suspends or terminates and plays described played multimedia file.
In a change case of the present embodiment, time model described in described step S117 is also based on equation below analysis：
Wherein, TtFor pushing the time；T is described played multimedia file broadcast start time；T1, T2, T3 and T4 are time constant.Preferably, described big data analysis system only is analyzed obtaining the described push time to described played multimedia file broadcast start time based on described time model.Preferably, described time point A includes described played multimedia file and terminates reproduction time.Preferably, described time point B includes described played multimedia file broadcast start time.Preferably, described T1=12, T2=15, T3=18, if described played multimedia file before 12 points or 15～18 points commence play out, judge described played multimedia file terminate play when to described client terminal push described information to be released；If described played multimedia file commences play out after 12～15 or 18 points, then judge to push described information to be released to described client terminal before described played multimedia file commences play out, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here, for example, user preengages in a video website at 14 points based on described client terminal and watches a film, then user can first browse some trailers related to described film before starting to watch described film, support brand advertising, film leading role to represent advertisement etc.；Again for example, user have purchased two 9 points of film ticket in a cinema, then user can receive food and drink entertainment advertisement of with lottery the vie to answer information related to described film, movie theatre periphery etc. after viewing terminates.Skilled artisan understands that, technical scheme described in this change case allows described big data analysis system to be based on described played multimedia file broadcast start time analysis and obtains the described push time, compared with step S117 described in above-mentioned embodiment illustrated in fig. 2, expand the range of described analysis method, if described big data analysis system cannot obtain described client terminal current time, remain to the described push time is analyzed, increased the applicable audient of described analysis method, improve the work efficiency of described big data analysis system.
In another change case of the present embodiment, described step S115 may alternatively be " obtaining the corresponding Internet registry information of described played multimedia file, multiple terminals that described Internet registry information is included are as described client terminal ".Specifically, described Internet registry information includes the corresponding history information of described client terminal and described client terminal corresponding user information.More specifically, described user profile includes identity information, preference information of filling in when user is registered on an internet site etc., the name of such as user, contact method, age, hobby, viewing type etc. based on described client terminal.Further, those skilled in the art may be referred to step S106 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Skilled artisan understands that, above-mentioned embodiment illustrated in fig. 2 can be understood as a change case of above-mentioned embodiment illustrated in fig. 1, in technical scheme described in above-mentioned embodiment illustrated in fig. 2, described big data analysis system is based on described time model and analyzes the described push time, determines the optimum push time by weighing the simple logic relation between described played multimedia file reproduction time and described client terminal current time；And above-mentioned embodiment illustrated in fig. 1 is then based on described time geography position model and analyzes the described push time, and place is play by relatively described client terminal current geographic position and described played multimedia file and judges that optimum releases news type, for example, if described played multimedia file broadcast start time or end reproduction time close on user's daily meal time, then user may receive the diet advertisement information of its current location periphery before described played multimedia file commences play out or after terminating to play, thus guiding user to go consumption in periphery Food Outlets；Again for example, if described played multimedia file broadcast start time is in the afternoon in the time period, user is likely to be received the amusement and leisure series advertisements information of described played multimedia file periphery, goes to while away the time before guiding user；Again for example, if user has just arrived in described played multimedia file before described played multimedia file commences play out and has play near place and leave described broadcasting place in the short period after described played multimedia file terminates to play, then judge that user plays place peripheral facility to described played multimedia file and do not have enough attention rates, preferably push the interactive information related to described played multimedia file to user, for example include information of vying to answer with lottery of described played multimedia file content etc., thus being reached for the purpose of different audients tailor-made customized information distribution scheme.Skilled artisan understands that, compared with above-mentioned embodiment illustrated in fig. 1, above-mentioned embodiment illustrated in fig. 2 allows described big data analysis system to remain to when obtaining described client terminal geographical location information judge the described push time based on described played multimedia document information, and it is preferably adapted for the situation that described client terminal does not open GPS location service with wireless internet connection or described client terminal；And above-mentioned embodiment illustrated in fig. 1 is preferably adapted for described client terminal and provides the situation of larger authority to described big data analysis system, if described big data analysis system can obtain the geographical position of described client terminal based on the communications protocol set up between described client terminal, the information such as the Internet historical record, then described big data analysis system just can be analyzed based on described time geography position model in the data basis comparing various dimensions, correspondingly its analysis result is also just more comprehensively, virtual condition closer to user, thus improve the specific aim of described information issue and by acceptance.
According to the third embodiment of the invention Fig. 3 illustrates, the flow chart of the big data analysis method based on viewing User Status, and described analysis result is used for releasing news, and described to be published information corresponding with described analysis result.Specifically, in the present embodiment, step S121 is first carried out, using lower for classification same in multiple for played multimedia file label informations frequency of occurrence highest N number of primary label information as information parameter, wherein N >=1.More specifically, the characteristic information of played multimedia file described in described label information, at least include performers and clerks' information of described played multimedia file, brand sponsored messages, prevailing scenario information etc., with described played multimedia file classify corresponding.Further, described information parameter includes one or more label informations of described played multimedia file, and it is used for searching corresponding information to be released in an information database to be released.Preferably, each described classification corresponds to one or more described label informations.Preferably, described big data analysis system counts described played multimedia file frame by frame and obtains the plurality of label information.In a preference，Described big data analysis system is sorted out to described label information during described played multimedia file is analyzed frame by frame，Count same classification each described label information occurrence number lower and using N number of label informations most for described occurrence number as described information parameter，Skilled artisan understands that，Described played multimedia file includes multiple classification，For example，" comedy "、" action "、" love " etc.，Then described big data analysis system is classified to the plurality of label information during analyzing described played multimedia file frame by frame，The label information that described played multimedia file is included is assigned under corresponding classification，Then described big data analysis system is based on multiple label information occurrence numbers under a certain classification of information publisher's demand statistics and obtains described information parameter，Described information publisher includes described information database provider to be released、Described big data analysis system background server priority instruction etc..In another preference, multiple label informations under classification multiple to described played multimedia file carry out frequency statistics to described big data analysis system respectively, obtain corresponding information parameter, the then corresponding information parameter set of each classification, described information parameter sets include one or more informations parameter, and described information parameter is corresponding with described label information.Preferably, described information parameter is stored in a storage device, and described storage device includes any one in flash memory, background thread, Flash, cloud storage device etc. or appoints multiple.Preferably, described N=10, those skilled in the art can also become according to actual needs and dissolve more embodiments, and this has no effect on the technology contents of the present invention.
Subsequently into the execution of step S122, corresponding multiple label information is searched in set U based on described information parameter.Specifically, described set U includes one or more information to be released and the corresponding label information of described information to be released.More specifically, described label information includes the principal character information of described information to be released, and for example, described information to be released corresponds to brand name, the corresponding mouthpiece's title of described information to be released or the corresponding descriptor of described information to be released etc..Further, those skilled in the art may be referred to step S112 described in above-mentioned embodiment illustrated in fig. 2, will not be described here.
Next execution step S123, based on natural language model, calculates the degree of association of described information parameter and described label information.Specifically, described label information includes described information principal character to be released, and for example, described information to be released corresponds to brand name, the corresponding mouthpiece's title of described information to be released or the corresponding descriptor of described information to be released etc..More specifically, described degree of association is based on equation below and calculates：
Wherein, x is information parameter, and X is label information, and N includes number of elements for described set U1, and n includes number of elements for described set U, and f function is based on natural language model and calculates acquisition.Further, those skilled in the art may be referred to step S104 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Then execution step S124, determines that the corresponding information to be released of described degree of association highest M label information is preferably to release news, wherein M >=1.Specifically, described correlation score is stored in described storage device, corresponding with described information to be released.More specifically, described big data analysis system preferably will be released news as described with described set U1 degree of association highest M information to be released in described set U.Further, those skilled in the art may be referred to step S105 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Next enter the execution of step S125, obtain the corresponding Internet registry information of described played multimedia file, multiple terminals that described Internet registry information is included are as described client terminal.Specifically, described Internet registry information includes the corresponding history information of described client terminal and described client terminal corresponding user information.More specifically, described user profile includes identity information, preference information of filling in when user is registered on an internet site etc., the name of such as user, contact method, age, hobby, viewing type etc. based on described client terminal.Further, those skilled in the art may be referred to step S106 described in the implemented example of above-mentioned Fig. 1, will not be described here.
Subsequently into the execution of step S126, obtain described client terminal and the status information of described played multimedia file respectively.Specifically, described status information at least includes any one in the history information of geographical location information, the temporal information of described client terminal or described client terminal of described client terminal or appoints multiple.More specifically, the status information of described played multimedia file includes any one in broadcasting place of the reproduction time of described played multimedia file, described played multimedia file etc. or appoints multiple.Further, those skilled in the art may be referred to step S107 described in above-mentioned embodiment illustrated in fig. 1, will not be described here.
Next execution step S127, analyzes described client terminal current time and described played multimedia file broadcast start time based on time model.Specifically, described client terminal current time is based on described client terminal background operating system and obtains.More specifically, described played multimedia file broadcast start time is based on described played multimedia paper document information acquisition.Further, those skilled in the art may be referred to step S117 described in above-mentioned embodiment illustrated in fig. 2, will not be described here.
Last execution step S128, determines the push time pushing described information to be released to described client terminal based on analysis result.Specifically, in described storage device, described storage device at least includes any one in flash memory, background thread, Flash, cloud storage device etc. or appoints multiple described information Store to be released.More specifically, described analysis result also is stored in described storage device.Further, those skilled in the art may be referred to step S118 described in above-mentioned embodiment illustrated in fig. 2, will not be described here.
In a change case of the present embodiment, described step S125 may alternatively be " using the terminal playing described played multimedia file as described client terminal ".Specifically, described client terminal is used for receiving and described preferably releases news.More specifically, described client terminal includes any one in mobile phone, IPAD, notebook computer, TV etc. or appoints multiple.Further, those skilled in the art may be referred to step S115 described in above-mentioned embodiment illustrated in fig. 2, will not be described here.
Skilled artisan understands that, step S121 described in the present embodiment can be understood as step S101 described in above-mentioned embodiment illustrated in fig. 1, a change case of step S111 described in described step S102 and described step S103 or above-mentioned embodiment illustrated in fig. 2, compared with above-mentioned embodiment illustrated in fig. 1 and above-mentioned embodiment illustrated in fig. 2, determine in technical scheme described in the present embodiment that the method for described information parameter is more easy, described big data analysis system is based on image identification system and data statistics system can be analyzed to described played multimedia file；And analysis method described in above-mentioned Fig. 1 and above-mentioned embodiment illustrated in fig. 2 is based on degree of association I and the positive screening more systematization to the plurality of label for the feature selection approach, excessively approximate or identical label information can be excluded, so that described information parameter to be related to field wider, it is the technical scheme that prior art does not adopt, the work efficiency of described big data analysis system, the operating experience of optimization information publisher can greatly be improved.
Fig. 4 illustrates according to the fourth embodiment of the invention, the structure chart of the big data analytical equipment based on viewing User Status, and described analysis result is used for releasing news, and described to be published information corresponding with described analysis result.Specifically, described analytical equipment 4 includes the first determination device 41, and it is used for determining information to be released based on the information parameter of played multimedia file；First acquisition device 42, it is used for obtaining the status information of the client terminal being associated with described played multimedia file；Second determination device 43, it is used for determining, based on described status information, the transmission time sending described information to be released.More specifically, described information parameter includes one or more label informations of described played multimedia file, and it is used for searching corresponding information to be released in an information database to be released.Further, described client terminal corresponding statess information at least includes any one in the history information of geographical location information, the temporal information of described client terminal or described client terminal of described client terminal or appoints multiple；Described played multimedia file corresponding statess information includes any one in broadcasting place of the reproduction time of described played multimedia file, described played multimedia file etc. or appoints multiple.Preferably, described analytical equipment 4 determines the described transmission time based on time geography bit model, determines the described transmission time by analyzing described client terminal corresponding statess information with described played multimedia file corresponding statess information.Preferably, described analytical equipment 4 is also based on time model and determines the described transmission time, determines the described transmission time by analyzing described client terminal current time with described played multimedia file broadcast start time.Preferably, described analytical equipment 4 correspondence analysis result is stored in a storage device, and described storage device includes any one in flash memory disk, background thread, Flash, cloud storage device etc. or appoints multiple.Preferably, described analysis result includes described information parameter, described information to be released, described information database to be released and described status information etc..
Preferably, described first determination device 41 includes the first processing meanss 411, its be used for using in multiple for described played multimedia file label informations with described played multimedia file dependency highest N number of primary label information as described information parameter, wherein N >=1；3rd determination device 413, it is used for determining corresponding information to be released in described set U based on described information parameter.Specifically, described label information includes the characteristic information of described played multimedia file, at least include performers and clerks' information of described played multimedia file, brand sponsored messages, prevailing scenario information etc., by acquisition is scanned frame by frame to described played multimedia file.More specifically, described set U includes one or more information to be released and the corresponding label information of described information to be released, and described label information includes the principal character information of described information to be released.Further, described N number of information parameter forms set U1, is stored in described storage device.In a preference，Described first processing meanss 411 weigh the correlation degree of multiple described label informations and described played multimedia file based on degree of association，Described label information is higher with described played multimedia file degree of association，Then association between the two is tightr，Described first processing meanss 411 will in described played multimedia file relation the most N number of described label information form described set U1，It is stored in described storage device，Then the described 3rd is called to determine that all information corresponding label information to be released are compared device 413 with described set U by the information parameter in described set U1，And the candidate of acquisition releases news and corresponding label information Store is in described storage device，Preferably，N=10，Those skilled in the art can also become according to actual needs and dissolve more embodiments，This has no effect on the technology contents of the present invention.
Preferably, described first processing meanss 411 include second processing device 4111, and it is used for calculating the degree of association I of the plurality of label information and a theme respectively；3rd processing meanss 4112, it is used for obtaining maximally related M label information, wherein M ＞ N with described theme based on described degree of association I；4th determination device 4113, it is used for selecting to determine that described N number of primary label information forms set U1 from described M label information based on positive feature selection approach.Specifically, described theme includes the feature general introduction in a certain respect of described played multimedia file, corresponding with described label information.More specifically, described label information can correspond to multiple described degree of association I, and each described degree of association is corresponding with a theme of described played multimedia file.Further, described forward direction feature selection approach is a greedy algorithm, and it is used for carrying out postsearch screening to described M label information, thus excluding excessively approximate useless label in described M label information.Preferably, described set U1 includes N number of label information of dependency minimum each other in described M label information.In a preference, described label information and described played multimedia file dependency are based on degree of association I and weigh, and described second processing device 4111 calculates described degree of association I by equation below：
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, and it is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) is the joint probability distribution of x and y.Described 3rd processing meanss 4112 are based on described degree of association I, and to described played multimedia file, all label informations arrange, preferably using M maximum for degree of association I numerical value described under same subject label information as described label information, thus preliminary screening is carried out to all label informations of described played multimedia file, then described 4th determination device 4113 is called to carry out postsearch screening to described M label information, N number of label information described in described set U1 is made to be provided simultaneously with two characteristics, described first characteristic includes described N number of label information degree of association I highest corresponding with described played multimedia file theme；It is minimum that described second characteristic includes described N number of label information degree of association I each other.Preferably, described set U1 is stored in described storage device.Preferably, M=20, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.
Preferably, described 4th determination device 4113 includes fourth process device 41131, and it is used for calculating the degree of association I1 of described M label information and described set U1 respectively；First choice device 41132, it is used for each label selecting described I1 minimum and is added in described set U1 for new label, and the number of labels up in described set U1 reaches N number of.Specifically, described degree of association I1 is based on equation below and calculates：
Wherein, x is label information to be screened in described M label information；yjFor j-th label information in described set U1, N is label information content in described set U1.More specifically, set U1 includes N number of information parameter with described played multimedia file degree of subject relativity highest, and it preferably chooses the minimum N number of label information composition of dependency each other from described M label information.In a preference, described 4th determination device 4113 extracts any one of described M label information label information first and puts into described set U1, then call described fourth process device 41131 that with the label information in described set U1, described M-1 label information is carried out degree of association I1 calculating respectively, being then based on described first choice device 41132 selects the label information that a label information minimum with label information correlation I1 in described set U1 is new to be added in described set U1, the like, until label information quantity reaches predetermined value N described in described set U1, the information parameter that then now the N number of described label information in described set U1 needs for described information publisher server, can be used for determining described information to be released.Preferably, N=10, those skilled in the art can also become according to actual needs and dissolve more embodiments, and this has no effect on the technology contents of the present invention.
Preferably, described 3rd determination device 413 includes the 6th processing meanss 4131, it is used for obtaining one or more described information to be released from described set U based on described theme, and described information to be released includes one or more label informations, and described label information is corresponding with described theme；7th processing meanss 4132, it is used for calculating the degree of association of one or more label informations that information parameter is included described in described set U1 with described information to be released based on natural language model；5th determination device 4133, it is used for determining that the described information to be released of described degree of association highest M is preferably to release news, wherein M >=1.Specifically, described information to be released corresponds to one or more described label informations.More specifically, described set U includes one or more described information to be released and described label information, and described information to be released is corresponding with described label information.Preferably, described information parameter obtains from described set U1, described 6th processing meanss 4131 are extracted an information parameter successively from described set U1 and call described 7th processing meanss 4132 to calculate the degree of association of described information parameter and multiple described label informations in described set U, finally obtain the degree of association of information to be released described in all informations parameter and described set U in described set U1.In a preference, described 7th processing meanss 4132 are based on equation below and calculate described degree of association：
Wherein, x is information parameter, and X is label information, and N includes number of elements for described set U1, and n includes number of elements for described set U, and f function is based on natural language model and calculates acquisition.Preferably, described preferably release news including coding information and described label information, described coding information includes the information write based on Javascript language, it is used for described preferably releasing news is converted into one section of code and is stored in described storage device, and decodable code is to be sent to client terminal after word, image or both combinations.
Preferably, described first acquisition device 42 includes the 6th determination device 421, and it is used for determining the client terminal being associated with described played multimedia file；Second acquisition device 422, it is used for obtaining the status information of described client terminal；3rd acquisition device 423, it is used for obtaining the status information of described played multimedia file.Specifically, described client terminal includes the terminal of currently playing described played multimedia file.More specifically, described client terminal also includes the one or more terminals obtaining based on Internet registry information.Further, described client terminal also includes showing the display terminal of described played multimedia file picture, and described display terminal is communicated based on wired or wireless way with described currently played multimedia file terminal of playing.Preferably, the status information of described client terminal at least includes any one in the history information of geographical location information, the temporal information of described client terminal or described client terminal of described client terminal or appoints multiple.Preferably, the status information of described played multimedia file includes any one in broadcasting place of the reproduction time of described played multimedia file, described played multimedia file etc. or appoints multiple.In a preference, described second acquisition device 422 obtains, based on GPS (global positioning system), the GPS information that described client terminal sends in real time so that it is determined that described client terminal current geographic position, and using described client terminal present system time as described client terminal current time, described second acquisition device 422 is by sending solicited message to described client terminal and receiving feedback information and obtain described client terminal status information, described solicited message is used for client terminal status information described in acquisition request, described feedback information is corresponding with described solicited message, including described client terminal status information；Described 3rd acquisition device 423 is then based on played multimedia file status information described in described played multimedia paper document information acquisition, described played multimedia paper document information includes described played multimedia file broadcast start time, file duration and plays place, then described 3rd acquisition device 423 sends solicited message to a background server and receives described feedback information, and described background server is corresponding with described played multimedia file and stores described played multimedia paper document information.Preferably, described status information is stored in described storage device.
Preferably, described 6th determination device 421 includes the 7th determination device 4211, and it is used for the terminal playing described played multimedia file as described client terminal.Specifically, described client terminal is used for receiving and described preferably releases news.More specifically, described client terminal includes any one in mobile phone, IPAD, notebook computer, TV etc. or appoints multiple.Further, described client terminal also includes and plays the display terminal that described played multimedia file terminal is connected.In a preference, user selects described display terminal as described client terminal, described display terminal is connected based on HDMI high-definition data line with a transferring device, described transferring device is connected based on WLAN with described playback terminal, then user can browse described played multimedia file on described display terminal.
Preferably, described second determination device 43 includes the 8th processing meanss 431, and it is used for calculating the degree of association F between the corresponding status information of described client terminal status information corresponding with described played multimedia file based on time geography position model；9th determination device 432, it is used for determining, based on described degree of association F, the push time pushing described information to be released to described client terminal.Specifically, the corresponding result of calculation of described degree of association F is stored in described storage device.Preferably, described degree of association F is based on equation below and calculates：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T is described played multimedia file broadcast start time；W is described client terminal current location；W is the broadcasting place of described played multimedia file.More specifically, described f function is used for indicating the degree of correlation between two variables, described f (t, T) is used for indicating the distance between described client terminal current time and described played multimedia file broadcast start time, is calculated based on equation below and obtains：
Wherein, described t and described T is based on 24 hours tabulating showing.Described t from described T more away from then described f numerical value bigger.For example, f (14,16)=0.17, f (15,15)=0.Preferably, described f (t, T) numerical value is between 0～1.Further, described f (w, W) is used for indicating that described client terminal current location is corresponding with described played multimedia file plays the distance between place, is calculated based on equation below and obtains：
Wherein, k is gauged distance, and for indicating, described client terminal current location is corresponding with described played multimedia file to play the distance between place to | w-W |.Preferably, more near then described f (w, the W) numerical value of described distance is less, and described f (w, W) value is also between 0～1.Preferably, the span of described degree of association F is also between 0～1.In a preference, described 9th determination device 432 presets first threshold θ to determine the described push time, if F is ＞ θ, then judge that the corresponding user of described client terminal occurs in described played multimedia file in the more early time and plays near place, using described client terminal current time as the described push time；If F is ＜ θ, then judge that the corresponding user of described client terminal closes on and just occur near described played multimedia file broadcasting place when described played multimedia file commences play out, described played multimedia file is terminated reproduction time as the described push time, after user terminates to watch described played multimedia file, send described information to be released to described client terminal again.
In a change case of the present embodiment, described first processing meanss 411 may alternatively be the 5th processing meanss 412, it is used for same classification time frequency of occurrence highest N number of primary label information in multiple for described played multimedia file label informations as described information parameter, wherein N >=1.Then described 5th processing meanss 412 are sorted out to described label information during described played multimedia file is analyzed frame by frame, count same classification each described label information occurrence number lower and using N number of label informations most for described occurrence number as described information parameter.In a preference, multiple label informations under classification multiple to described played multimedia file carry out frequency statistics to described 5th processing meanss 412 respectively, obtain corresponding information parameter, the then corresponding information parameter set of each classification, described information parameter sets include one or more informations parameter, and described information parameter is corresponding with described label information.
In another change case of the present embodiment, described 7th determination device 4211 may alternatively be the 8th determination device 4212, it is used for obtaining the corresponding Internet registry information of described played multimedia file, and multiple terminals that described Internet registry information is included are as described client terminal.Specifically, described Internet registry information includes the corresponding history information of described client terminal and described client terminal corresponding user information.More specifically, described Internet registry information also includes Transaction Information on described internet site for the nearest n times of described client terminal.Preferably, described Transaction Information includes the Ticketing information that described client terminal is produced based on register account number, described register account number is corresponding with described internet electronic ticket-booking system, and described Ticketing information includes described played multimedia file and corresponds to title, reproduction time, plays place etc..Skilled artisan understands that, described 8th determination device 4212 is communicated with described internet site in real time, thus obtaining described Internet registry information in time, and nearest n times are completed the one or more terminals concluded the business as described client terminal in described internet site.Preferably, N=1, those skilled in the art can also become according to actual needs and dissolve more embodiments, will not be described here.
In another change case of the present embodiment, described 8th processing meanss 431 may alternatively be the first analytical equipment 433, and it is used for analyzing described client terminal current time and described played multimedia file broadcast start time based on time model；Correspondingly, described 9th determination device 432 may alternatively be the tenth determination device 434, and it is used for determining, based on analysis result, the push time pushing described information to be released to described client terminal.Specifically, described time model is analyzed based on equation below：
Wherein, Tt is the push time；T is described played multimedia file broadcast start time；T is described client terminal current time；β is constant.More specifically, the described push time is based on described time point A and described time point B and determines.In a preference, described β=0, then described first analytical equipment 433 judges the precedence relationship between described played multimedia file broadcast start time and described client terminal current time, if it is t-T ＞ 0 that described played multimedia file broadcast start time is later than described client terminal current time, judge in A time point to described client terminal pushed information；If described played multimedia file broadcast start time early than or be t-T≤0 equal to described client terminal current time, judge in B time point to described client terminal pushed information.In another preference, described time model is also based on equation below analysis：
Wherein, TtFor pushing the time；T is described played multimedia file broadcast start time；T1, T2, T3 and T4 are time constant.Preferably, described time point A includes described played multimedia file and terminates reproduction time.Preferably, described time point B includes described played multimedia file broadcast start time.Preferably, described T1=12, T2=15, T3=18, if the analysis result of described first analytical equipment 433 show described played multimedia file before 12 points or 15～18 points commence play out, described tenth determination device 434 judge described played multimedia file terminate play when to described client terminal push described information to be released；If the analysis result of described first analytical equipment 433 shows that described played multimedia file commences play out after 12～15 or at 18 points, described tenth determination device 434 judges to push described information to be released to described client terminal before described played multimedia file commences play out.
Fig. 5 shows the network topology structure schematic diagram of typical application scenarios according to the present invention, wherein said analysis system is communicated with analytical equipment described in above-mentioned embodiment illustrated in fig. 4, specifically, in the present embodiment, described analysis system includes first server, it is used for determining information to be released based on the information parameter of played multimedia file, obtains the status information of client terminal being associated with described played multimedia file and determines the transmission time sending described information to be released based on described status information.More specifically, described analysis system also includes second server, and it is used for determining the client terminal being associated with described played multimedia file.In application scenarios，For a multimedia file is opened based on a terminal，Described terminal background operating system sends the first solicited message to described first server，Described first solicited message includes described played multimedia paper document information，After then described first server receives described played multimedia paper document information，It is based preferably on degree of association I and positive feature selection approach obtains described information parameter，And search one or more described information to be released from an information database to be released，Described information to be released is corresponding with described information parameter，Then described first server sends the second solicited message to described second server，Described second solicited message includes described playback terminal information，Then described second server based on described playback terminal information using the terminal of currently playing described played multimedia file as after described client terminal，Send the second feedback information to described first server，Described second feedback information is corresponding with described second solicited message，Including described client terminal status information，Then described first server is based on described second feedback information and described played multimedia paper document information determines the described transmission time，And the first feedback information is sent to described client terminal based on the described transmission time，Described first feedback information is corresponding with described first solicited message，Including described information to be released，Described client terminal after receiving described first feedback information, described information to be released is illustrated on described client terminal current active window and browses for user.In another application scenarios, described second server is based on the corresponding Internet registry information of described played multimedia file and obtains one or more end messages, and using described terminal as described client terminal.Preferably, described first server is based on time geography position model and determines the described transmission time, if described client terminal occurs in described played multimedia file before described played multimedia file commences play out and plays near place, using described client terminal current time as the described transmission time.Preferably, described first server is also based on time model and determines the described transmission time, if described played multimedia file broadcast start time early than or be equal to described client terminal current time, described played multimedia file is terminated reproduction time as the described transmission time.In another application scenarios, described analysis system also includes the 3rd server, it is used for including information delivery server to the 3rd server described in the described client terminal described information to be released of transmission, preferably, described first server sends solicited message to described 3rd server, described solicited message includes described information to be released, described client terminal information and described transmission time, then send described information to be released based on the described transmission time to described client terminal after described 3rd server receives described solicited message.Preferably, described information database to be released is stored in described 3rd server, then described first server sends solicited message to the 3rd server after obtaining described information parameter, described solicited message includes described information parameter, described 3rd server receives the degree of association analyzing described information parameter and one or more information to be released in described information database to be released after described solicited message based on degree of association I1, thus obtaining one or more information to be released as preferably releasing news, finally described 3rd server is to described first server send feedback information, described feedback information is corresponding with described solicited message, preferably release news including one or more of.Skilled artisan understands that，In order to express easily，The present embodiment will be separately shown to described first server and second server，Actually they are the concept of a logic，In actual applications，Described first server and described second server can also be bound together，It is located in same server，Or combine two-by-two，Or respective individualism，Correspondingly，Described first server and described client terminal are also to be combined together，Then described first server is preset in described client terminal，After user opens a multimedia file on described client terminal，Described client terminal operating system calls described first server that described played multimedia file is analyzed，And then it is based on information to be released described in described played multimedia document information parameter determination，And the described transmission time is determined based on described client terminal status information and described played multimedia file status information，Thus sending described information to be released in the described transmission time to described client terminal.
Above the specific embodiment of the present invention is described.It is to be appreciated that the invention is not limited in above-mentioned particular implementation, those skilled in the art can make various modifications or modification within the scope of the claims, and this has no effect on the flesh and blood of the present invention.

Claims (28)

1. a kind of big data analysis method based on viewing User Status, described analysis result is used for Release news, described be published information corresponding with described analysis result it is characterised in that include Following steps：
A. information to be released is determined based on the information parameter of played multimedia file；
B. obtain the status information of the client terminal being associated with described played multimedia file；
C. the transmission time sending described information to be released is determined based on described status information.
2. analysis method according to claim 1 is it is characterised in that multiple described to be released Information forms a set U, and described step a comprises the steps：
A1. by multiple for described played multimedia file label informations with described played multimedia Correlation of files highest N number of primary label information is as described information parameter, wherein N >=1；
A2. corresponding information to be released is determined in described set U based on described information parameter.
3. analysis method according to claim 2 it is characterised in that described set U also Including the corresponding label information of described information to be released.
4. the analysis method according to Claims 2 or 3 is it is characterised in that described step A1 comprises the steps：
A11. the degree of association I of the plurality of label information and a theme, wherein said correlation are calculated respectively Degree I passes through following mutual information formula and calculates acquisition：
I [ x , y ] ≡ K L ( p ( x , y ) | | p ( x ) p ( y ) ) = - ∫ ∫ p ( x , y ) ln ( p ( x ) p ( y ) p ( x , y ) ) d x d y
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, It is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) be x and The joint probability distribution of y；
A12. it is based on described degree of association I and obtain the maximally related M label information with described theme, its Middle M ＞ N；
A13. it is based on positive feature selection approach to select to determine described N from described M label information Individual primary label information forms set U1.
5. analysis method according to claim 4 is it is characterised in that described forward direction feature is selected Selection method is a greedy algorithm, and described step a13 comprises the steps：
- to calculate described M label information respectively based on equation below related to described set U1 Degree I1：
I 1 ( x , y ) = Σ j = 1 N I ( x , y j )
Wherein, x is label information to be screened in described M label information；yjFor described set U1 In j-th label information；
- select a minimum label of described I1 to be added in described set U1 for new label every time, Number of labels up in described set U1 reaches N number of.
6. the analysis method according to Claims 2 or 3 is it is characterised in that described step a Comprise the steps：
A1 '. by lower for classification same in multiple for described played multimedia file label informations frequency of occurrence Highest N number of primary label information is as described information parameter, wherein N >=1；
A2. corresponding information to be released is determined in described set U based on described information parameter.
7. the analysis method according to any one of claim 2 to 5 is it is characterised in that institute State step a2 to comprise the steps：
A21. one or more described letters to be released are obtained from described set U based on described theme Breath, described information to be released includes one or more label informations, described label information and described master Topic is corresponding；
A22. it is based on natural language model, information described in described set U1 is calculated by equation below The degree of association of one or more label informations that parameter is included with described information to be released：
F ( x , X ) = Σ i = 1 N Σ j = 1 n f ( x i , X j )
Wherein, x is information parameter；X is label information；
A23. determine that the described information to be released of described degree of association highest M is preferably to release news, Wherein M >=1.
8. analysis method according to any one of claim 1 to 6 is it is characterised in that institute State step b to comprise the steps：
B1. determine the client terminal being associated with described played multimedia file；
B2. obtain the status information of described client terminal, wherein said status information at least includes as follows In information any one or appoint multiple：
The geographical location information of-described client terminal；
The temporal information of-described client terminal；
The history information of-described client terminal.
9. analysis method according to claim 7 is it is characterised in that described step b is also wrapped Include following steps：
B3. obtain the status information of described played multimedia file, wherein said status information is at least Including any one in following information or appoint multiple：
The reproduction time of-described played multimedia file；
The broadcasting place of-described played multimedia file.
10. the analysis method according to claim 7 or 8 is it is characterised in that described step B1 comprises the steps：
B11. using the terminal playing described played multimedia file as described client terminal.
11. analysis methods according to claim 7 or 8 are it is characterised in that described step B1 also comprises the steps：
B11 '. obtain the corresponding Internet registry information of described played multimedia file, will be described mutual Multiple terminals that networking log-on message includes are as described client terminal.
12. analysis methods according to any one of claim 1 to 10 it is characterised in that Described step c comprises the steps：
C1. it is based on time geography position model and calculate the corresponding status information of described client terminal and institute State the degree of association between the corresponding status information of played multimedia file, wherein said degree of association leads to Cross equation below to calculate：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T starts for described played multimedia file Reproduction time；W is described client terminal current location；W is described played multimedia file Play place；
C2. determined based on described degree of association F and push described information to be released to described client terminal The push time.
13. according to claim 12 analysis method it is characterised in that described f (t, T) be based on such as Lower formula calculates：
f ( t , T ) = | t - T | 24
Wherein, described t and described T is based on 24 hours tabulating showing.
Correspondingly, described f (w, W) is based on equation below and obtains：
f ( w , W ) = 1 - | w - W | K
Wherein, k be gauged distance, | w-W | for indicate described client terminal current location with described
Played multimedia file is corresponding to play the distance between place.
14. analysis methods according to any one of claim 1 to 10 it is characterised in that Described step c also comprises the steps：
C1 '. described client terminal current time and described played multimedia are analyzed based on time model File broadcast start time；
C2 '. the push pushing described information to be released to described client terminal is determined based on analysis result Time.
A kind of 15. big data analytical equipments based on viewing User Status, described analysis result by with In releasing news, described be published information corresponding with described analysis result it is characterised in that bag Include：
First determination device, it is pending for being determined based on the information parameter of played multimedia file Cloth information；
First acquisition device, it is used for obtaining the client being associated with described played multimedia file The status information of terminal；
Second determination device, it is used for determining the described information to be released of transmission based on described status information The transmission time.
16. analytical equipments according to claim 15 are it is characterised in that multiple described pending Cloth information forms a set U, and described first determination device includes：
First processing meanss, its be used in multiple for described played multimedia file label informations and The N number of primary label information of described played multimedia file dependency highest is joined as described information Number, wherein N >=1；
3rd determination device, it is used for determining correspondence in described set U based on described information parameter Information to be released.
17. analytical equipments according to claim 16 are it is characterised in that described first is processed Device includes：
Second processing device, it is related to a theme for calculating the plurality of label information respectively Degree I, wherein said degree of association I pass through following mutual information formula and calculate acquisition：
I [ x , y ] ≡ K L ( p ( x , y ) | | p ( x ) p ( y ) ) = - ∫ ∫ p ( x , y ) ln ( p ( x ) p ( y ) p ( x , y ) ) d x d y
Wherein, x is label information to be screened；Label information on the basis of y；KL function is relative entropy, It is used for representing the relation between two distributions；P function is probability-distribution function；P (x, y) be x and The joint probability distribution of y；
3rd processing meanss, it is used for maximally related with described theme based on described degree of association I acquisition M label information, wherein M ＞ N；
4th determination device, it is used for based on positive feature selection approach from described M label information Middle selection determines that described N number of primary label information forms set U1.
18. analytical equipments according to claim 17 it is characterised in that described forward direction feature System of selection is a greedy algorithm, and described 4th determination device includes：
Fourth process device, its be used for based on equation below calculate respectively described M label information and The degree of association I1 of described set U1：
I 1 ( x , y ) = Σ j = 1 N I ( x , y j )
Wherein, x is label information to be screened in described M label information；yjFor described set U1 In j-th label information；
First choice device, it is used for each label selecting described I1 minimum and adds for new label Enter in described set U1, the number of labels up in described set U1 reaches N number of.
19. analytical equipments according to claim 16 are it is characterised in that described first determines Device also includes：
5th processing meanss, its be used for by multiple for described played multimedia file label informations together The lower frequency of occurrence highest N number of primary label information of one classification is as described information parameter, wherein N ≥1；
3rd determination device, it is used for determining correspondence in described set U based on described information parameter Information to be released.
20. analytical equipments according to any one of claim 16 to 18 it is characterised in that Described 3rd determination device includes：
6th processing meanss, it is used for obtaining one or many based on described theme from described set U Individual described information to be released, described information to be released includes one or more label informations, described mark Label information is corresponding with described theme；
7th processing meanss, it is used for, based on natural language model, calculating described by equation below One or more label informations that information parameter described in set U1 is included with described information to be released Degree of association：
F ( x , X ) = Σ i = 1 N Σ j = 1 n f ( x i , X j )
Wherein, x is information parameter；X is label information；
5th determination device, it is used for determining the described information to be released of described degree of association highest M For preferably releasing news, wherein M >=1.
21. analytical equipments according to any one of claim 15 to 19 it is characterised in that Described first acquisition device includes：
6th determination device, it is used for determining the client being associated with described played multimedia file Terminal；
Second acquisition device, it is used for obtaining the status information of described client terminal, wherein said shape State information at least includes any one in following information or appoints multiple：
The geographical location information of-described client terminal；
The temporal information of-described client terminal；
The history information of-described client terminal.
22. analytical equipments according to claim 21 are it is characterised in that described first obtains Device also includes：
3rd acquisition device, it is used for obtaining the status information of described played multimedia file, its Described in status information at least include in following information any one or appoint multiple：
The reproduction time of-described played multimedia file；
The broadcasting place of-described played multimedia file.
23. analytical equipments according to claim 22 are it is characterised in that the described 6th determines Device includes：
7th determination device, it is used for the terminal playing described played multimedia file as institute State client terminal.
24. analytical equipments according to claim 22 are it is characterised in that the described 6th determines Device also includes：
8th determination device, it is used for obtaining described played multimedia file corresponding the Internet note Volume information, multiple terminals that described Internet registry information is included are as described client terminal.
25. analytical equipments according to any one of claim 15 to 24 it is characterised in that Described second determination device includes：
8th processing meanss, it is used for calculating described client terminal pair based on time geography position model Degree of association between the status information answered status information corresponding with described played multimedia file, Wherein said degree of association is calculated by equation below：
F=f (t, T) * f (w, W)
Wherein, t is described client terminal current time；T starts for described played multimedia file Reproduction time；W is described client terminal current location；W is described played multimedia file Play place；
9th determination device, it is used for determining to the push of described client terminal based on described degree of association F The push time of described information to be released.
26. according to claim 26 analytical equipment it is characterised in that described f (t, T) be based on such as Lower formula calculates：
f ( t , T ) = | t - T | 24
Wherein, described t and described T is based on 24 hours tabulating showing.
Correspondingly, described f (w, W) is based on equation below and obtains：
f ( w , W ) = 1 - | w - W | K
Wherein, k be gauged distance, | w-W | for indicate described client terminal current location with described
Played multimedia file is corresponding to play the distance between place.
27. analytical equipments according to any one of claim 15 to 24 it is characterised in that Described second determination device also includes：
First analytical equipment, its be used for based on time model analyze described client terminal current time with Described played multimedia file broadcast start time；
Tenth determination device, it is used for determining to described in the push of described client terminal based on analysis result The push time of information to be released.
A kind of 28. big data analysis systems based on viewing User Status, described analysis result by with In releasing news, described be published information corresponding with described analysis result it is characterised in that bag Include：
First server, it is to be released for being determined based on the information parameter of played multimedia file Information, obtains the status information of client terminal being associated with described played multimedia file base Determine the transmission time sending described information to be released in described status information；
Second server, it is used for determining the client being associated with described played multimedia file eventually End；
Described analysis system also includes the analysis dress according to any one of claim 15～27 Put, wherein, described first server, second server and described analytical equipment communicate.
CN201510507056.0A 2015-08-18 2015-08-18 A kind of big data analysis method based on viewing User Status, apparatus and system Pending CN106469159A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510507056.0A CN106469159A (en) 2015-08-18 2015-08-18 A kind of big data analysis method based on viewing User Status, apparatus and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510507056.0A CN106469159A (en) 2015-08-18 2015-08-18 A kind of big data analysis method based on viewing User Status, apparatus and system

Publications (1)

Publication Number Publication Date
CN106469159A true CN106469159A (en) 2017-03-01

Family

ID=58214232

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510507056.0A Pending CN106469159A (en) 2015-08-18 2015-08-18 A kind of big data analysis method based on viewing User Status, apparatus and system

Country Status (1)

Country Link
CN (1) CN106469159A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112597653A (en) * 2020-12-24 2021-04-02 南京城建隧桥经营管理有限责任公司 Method for determining information display duration of outdoor multimedia platform
CN114422835A (en) * 2021-12-29 2022-04-29 上海数即数据科技有限公司 Advertisement directional promotion platform based on big data analysis
CN117557317A (en) * 2023-12-14 2024-02-13 广州睿自游信息科技有限公司 Scene recommendation method and system based on ticket buying record

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1852109A (en) * 2005-12-31 2006-10-25 华为技术有限公司 Information issurance system and method
CN103368898A (en) * 2012-03-26 2013-10-23 中兴通讯股份有限公司 Method and system for accomplishing information push
CN105243103A (en) * 2015-09-19 2016-01-13 杭州电子科技大学 Content based push time determination method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1852109A (en) * 2005-12-31 2006-10-25 华为技术有限公司 Information issurance system and method
CN103368898A (en) * 2012-03-26 2013-10-23 中兴通讯股份有限公司 Method and system for accomplishing information push
CN105243103A (en) * 2015-09-19 2016-01-13 杭州电子科技大学 Content based push time determination method

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112597653A (en) * 2020-12-24 2021-04-02 南京城建隧桥经营管理有限责任公司 Method for determining information display duration of outdoor multimedia platform
CN112597653B (en) * 2020-12-24 2024-04-19 南京城建隧桥智慧管理有限公司 Outdoor multimedia platform information display duration determining method
CN114422835A (en) * 2021-12-29 2022-04-29 上海数即数据科技有限公司 Advertisement directional promotion platform based on big data analysis
CN117557317A (en) * 2023-12-14 2024-02-13 广州睿自游信息科技有限公司 Scene recommendation method and system based on ticket buying record
CN117557317B (en) * 2023-12-14 2024-04-16 广州睿自游信息科技有限公司 Scene recommendation method and system based on ticket buying record

Similar Documents

Publication Publication Date Title
US20170303010A1 (en) Methods and apparatus for enhancing a digital content experience
US9111286B2 (en) Multiple actions and icons for mobile advertising
CN107430618B (en) System and method for enabling user voice interaction with a host computing device
US20180039621A1 (en) Context analysis for message enhancement
CN101178737B (en) Method and system for publishing information related to internet key character
WO2017092244A1 (en) Information processing method, client and computer storage medium
US20110251902A1 (en) Target Area Based Content and Stream Monetization Using Feedback
US20080275764A1 (en) Content management system and method
US20120304223A1 (en) Ad selection and next video recommendation in a video streaming system exclusive of user identity-based parameter
US20100049608A1 (en) Third party content management system and method
US20090228361A1 (en) Cognitive scheduler for mobile platforms
CN102027498A (en) Critical mass billboard
CN102132588A (en) Platform for mobile advertising and micro-targeted promotions
JP2013519162A (en) Integrated advertising system
JP2012519926A (en) Targeting by context information of content using monetization platform
US20090164310A1 (en) Method for providing mobile commerce and revenue optimization
KR20160036518A (en) Selectable Styles for Text Messaging System User Devices
CN101599152A (en) A kind of device and method according to the user customized information issuing advertisement information
CN106415644A (en) Dynamic content item creation
US20080305781A1 (en) Cognitive scheduler
KR20160036522A (en) Selectable Styles for Text Messaging System Font Service Providers
EP3001379A1 (en) Selectable text messaging styles for brand owners
CN106407220A (en) Information publishment control method, control apparatus, and system
KR20160036520A (en) Selectable Styles for Text Messaging System Publishers
CN101425160A (en) Advertisement precision marketing system by utilizing modern media communications technology

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20170301