CN104935963A - Video recommendation method based on timing sequence data mining - Google Patents
Video recommendation method based on timing sequence data mining Download PDFInfo
- Publication number
- CN104935963A CN104935963A CN201510290170.2A CN201510290170A CN104935963A CN 104935963 A CN104935963 A CN 104935963A CN 201510290170 A CN201510290170 A CN 201510290170A CN 104935963 A CN104935963 A CN 104935963A
- Authority
- CN
- China
- Prior art keywords
- user
- interest
- project
- video
- scoring
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/258—Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
- H04N21/25866—Management of end-user data
- H04N21/25891—Management of end-user data being end-user preferences
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4667—Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies
Abstract
The invention relates to a video recommendation method based on timing sequence data mining. The method comprises the steps that 1) video interest gradient change of a user is analyzed via third party data and a user interest gradient curve is obtained, and the singular points of the user interest gradient curve act as time points of user interest migration; 2) the latest interest migration time points of the user are confirmed, and user-item scores after the latest interest migration time points of the user are acquired so that a user-item score matrix meeting the current interest of the user within a selected time window is established; and 3) user customized video recommendation is performed by using a random walk model on the basis of the user-item score matrix. An interest migration problem in the customized video recommendation is considered, and customized video recommendation is performed though integration of a time window method and on the basis of the Random Walker trust degree model so that video recommendation accuracy and efficiency are enhanced.
Description
Technical field
The invention belongs to video technique, video recommendations technical field, be specifically related to a kind of video recommendation method based on timing driving.
Background technology
Along with the development of the communication technology, network availability bandwidth increases sharply, viewing video for pc user even mobile phone users be all a lax usual thing, this development also bringing internet video industry like a raging fire, and video recommendation system is the important component part in internet video service.Initial video recommendation system is recommended according to the similitude between film, and two users for viewing same portion film can provide identical recommendation results.But in recent years along with the successful Application of personalized recommendation method on search engine and shopping website, video recommendations field is also more and more paid attention to recommending for the individualized video of user characteristics.Individualized video is recommended to consider the information such as the viewing history of user, social networks and current broadcasting film, recommends this user may interested film.Can predict, personalized recommendation method also has broad application prospects gathering around and the expection of considerable profit at video field.
Individualized video recommend method is roughly divided into content filtering method (Content-based Recommendation) and collaborative filtering method (Collaborative Filtering Recommendation) two classes.The film information that content filtering method is liked according to user, recommend other similar in terms of content videos to user, such as party A-subscriber likes " allowing bullet fly ", and so content filtering system can recommend Jiang Wen as other works of director as " one step away ".The essence of content filtering method is the similarity calculating film as can be seen here, and generally, content filtering method first defines a series of label for characterizing a video features and user preferences, thus video similarity problem is converted into vector distance computational problem.Collaborative filtering method utilizes user to calculate user's similitude or item similarity to the history score data of project, and then recommends according to similitude.The all users of collaborative filtering method record are to the scoring of all items, and form the user-project rating matrix of a M*N dimension, wherein M is number of users, and N is item number, a unit R in matrix
m,nrepresent that user m is to the scoring of project n, in matrix, every a line represents a user vector, can obtain the similarity between user vector by similarity calculating method, the similarity also namely between user; Each row in matrix represent a project, the same similarity that can obtain between project with similarity calculating method.According to the similarity between user, recommend the project having the user of similar interests to like with him can to a user, this recommendation is called the collaborative filtering (UserCF) based on user; According to the similarity between project, recommend the project similar to the project that he likes can to a user, this recommendation is called project-based collaborative filtering (ItemCF).
No matter information filtering algorithm or collaborative filtering is all the angle from user interest, and for user recommends its film that may like, but in actual life, the interest of user gradually changes along with the time often, and this phenomenon is called interest migration.If do not consider the interest migration of user when carrying out individualized video to user and recommending, just mean system only may recommend him to user before the film liked, but the interest nearest with it is not inconsistent, and this reduces the accuracy of recommendation.
Summary of the invention
The present invention is directed to the user interest migration problem in individualized video commending system, a kind of video recommendation method based on timing driving is provided, the accuracy and efficiency of video recommendations can be improved.
The technical solution used in the present invention is as follows:
Based on a video recommendation method for timing driving, its step comprises:
1) by the interest graded of third party's data analysis user to video, and user interest gradient curve is obtained, using the time point that the singular point of user interest gradient curve moves as user interest;
2) determine the interest transit time point that user is nearest, the user-project scoring later to the interest transit time point that user is nearest gathers, thus sets up the user the meeting user's current interest-project rating matrix in window seclected time;
3) based on described user-project rating matrix, recommended by the individualized video using random walk model to carry out user.
Further, step 1) described third party's data are preferably bean cotyledon scoring, draw user interest gradient curve by the bean cotyledon scoring history analyzing user.
Further, step 1) method that obtains described user interest gradient curve is: each attribute of video is formed an attribute vector; At each time point, each attribute of video is weighted according to current time and the user of time before record of marking, obtains a user interest vector; According to a series of user interests vector of user's multiple time point on a timeline, analyze and obtain time dependent user interest gradient curve.Each attribute of described video comprises: title, performer, director, age, brief introduction etc.
Further, step 2) described user-project scoring comprises explicit scoring and implicit scores; Described explicit scoring is when user watches video, the scoring that user carries out the current video watched; Described implicit scores be recording user viewing, redirect, collect, share action, according to the different action prediction users of user to the fancy grade of video.
Further, step 2) by quantizing and normalization process user-project scoring, obtain interest transit time later user-project rating matrix, the every a line in matrix represents a user characteristics, and each row represents an item characteristic.
Further, step 3) first calculate user u according to user-project rating matrix and be in the transition probability of project j in kth step, then calculate user u to the overall probability of project j, then calculate the overall probability of all users to all items, thus obtain user-entry sorting matrix
individualized video recommendation is carried out according to this matrix.
Key point of the present invention is: by third party's data to user modeling, finds user interest migration; By determining the interest transit time point that user is nearest, set up user-project rating matrix; Collaborative filtering method is replaced to carry out individualized video recommendation by using random walk model.
The present invention is directed to the user interest migration problem in individualized video commending system, use third party's data analysis user interest graded, find user interest transit time, thus determine suitable training time window size; In time window, create trust metric model based on Random Walker random algorithm, thus realize recommending for the individualized video of certain user.The method considers the interest migration problem in individualized video recommendation, and has merged time window method and carried out individualized video recommendation based on the trust metric model of Random Walker, improves the accuracy and efficiency of video recommendations.
Accompanying drawing explanation
Fig. 1 is the video recommendation method overview flow chart based on timing driving.
Fig. 2 is the user interest migration detection curve based on interest gradient analysis.
Embodiment
For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and below by specific embodiments and the drawings, the present invention will be further described.
The present invention utilizes third party's data (such as utilizing bean cotyledon to mark) to find user interest migration, determine training data time window size, thus the impact of shielding user interest migration, and then the trust metric model based on Random Walker is utilized to carry out individualized video recommendation.Fig. 1 is the overview flow chart of the inventive method.
1. the user interest migration based on third party's data finds
As previously mentioned, windowhood method overcame user interest and moved the impact caused recommendation results service time of the present invention.The time window of fixed size can not the interest migration of accurate response user, therefore, in order to obtain suitable time window size, the present invention uses third party's data analysis user interest graded, estimating subscriber's interest transit time, thus determine the size of time window.
Recommend owing to the present invention is directed to video resource, and in the network environment of China, bean cotyledon scoring is best suited for the third party website of objective evaluation user video interest undoubtedly.Therefore, the present invention analyzes the bean cotyledon scoring history of user, draws user interest gradient curve, and wherein nearest singular point is moved initial mark as the last interest of user, determines the size of time window in video recommendations model with this.
1) bean cotyledon user data modeling
Bean cotyledon is that the film of largest domestic is shared and comments on platform, more than the registration amount of ten million user, bean cotyledon is marked and enjoys very large influence power in Video service industry.User on bean cotyledon website marks and adopts very full marks system, therefore eliminates data normalization process, can directly the history scoring record of user on bean cotyledon be brought as training data.
2) user interest modeling
In order to be described film, the title of film, performer, director, age, brief introduction etc. as a vector, are represented a film with this attribute vector by the present invention.In such as table 1, film a just can represent with [Name a, Actor a Actor b, Director a, 2014, Keyword a Keyword d] such attribute vector.
According to the scoring record of certain user to its viewing film, the attribute of each film can be weighted by scoring, thus obtain the hobby characteristic vector of this user.Such as certain user A viewed 4 films altogether, be respectively film a, film b, film c, film d, and the scoring of A to four films are recorded as shown in table 1.
Table 1 user viewing history and scoring record sheet
So, the interest vector of this user just can be expressed as I (t)=[name a*9+name b*5+name c*9+named*9, actor a*24+actor b*9+actor c*14+actor d*18, director a*18+director b*9+directorc*9, 2014*27+2010*5, keyword a*18+keyword b*9+keyword c*9+keyword e*9+keyword f*9+keyword g*5+keyword h*5]=[name a, actor a, director a, 2014, keyword a] (assuming that each only gets a value by weight sequencing in interest vector, what weights were identical gets a value at random).
As time goes on, the scoring record of user can get more and more, therefore at each time point, we can according to current time and before the time history scoring record, obtain a user interest vector, accordingly, we can obtain the interest vector of this user multiple time point on a timeline, and each interest vector represents the preference of user at corresponding time point viewing video.
3) user interest migration finds
According to above-mentioned a series of user interest vectors, can analyze and obtain time dependent user interest gradient, find the time point that the singular point of gradient curve moves as interest.This point is the mark of user interest migration, thus provides the foundation of time window size for follow-up collaborative filtering video recommendations algorithm.
The interest gradient calculation formula of certain time point t is:
D(t)=I(t)-I(t-1)
Wherein, D (t) the user interest gradient that is t; The user interest vector that I (t) is t; The user interest vector that I (t-1) is t previous moment.Conveniently calculate, the rule in regulation subtraction is, subtracts each other if identical, is 0, if difference is subtracted each other, is 1, embodies the otherness between two vectors with this.Such as [name a, actor a, director a, 2014, keyword a]-[name b, actor a, director c, 2014, keyword a]=[1,0,1,0,0].As can be seen here, the modulus value of user interest gradient vector is larger, just represents interests change more obvious.
Figure 2 shows that interest gradient, wherein transverse axis is the time, and the longitudinal axis is the modulus value of interest gradient.Can see from this figure, the interest of user there occurs an interest migration respectively in February, 2009, in December, 2009, in June, 2012 and in June, 2014.
2. user-project rating matrix
Find result according to above-mentioned interest migration, the interest transit time point that user is nearest can be determined.It is nonsensical for doing video recommendations according to the user items scoring record before this time point, because that does not meet the up-to-date interest of user.Therefore user-project scoring that the present invention is only later to interest transit time point gathers, thus obtains the user-project rating matrix meeting each user's current interest most.
1) user-project scoring gathers
In video recommendation system, explicit scoring and implicit scores two kinds of users-project scoring acquisition mode can be had.Wherein, explicit scoring is when user watches video, allows user to mark (being generally 5 points of full marks or 10 points of full marks systems) to the current video watched; Implicit scores be recording user viewing, redirect, collect, the action such as to share, according to the different action prediction users of user to the fancy grade of this video.User of the present invention-project scoring takes into account explicit scoring and implicit scores two kinds scoring acquisition mode, thus reacts user more all-sidedly and accurately to the fancy grade of each video.
2) user-project scoring normalization
In the present invention, explicit scoring refer to user watch in video process to watch that film makes score, conveniently subsequent calculations, normalizes to full marks 1 point by all explicit scorings.Implicit scores comprise user viewing, redirect, collect, the action such as to share, suppose in the present invention that score value corresponding to the action of user is as shown in table 2.
The user concealed scoring of table 2 quantizes and normalization table
Action | Share | Collection | Watch and watch ending | Viewing is midway redirect also |
Scoring | 10 | 9 | 8 | 5 |
Normalization is marked | 1 | 0.9 | 0.8 | 0.5 |
Through above-mentioned quantification and normalization process, can obtain interest transit time later user-project rating matrix, the every a line in matrix represents a user characteristics, and each row represents an item characteristic.
3. based on the video recommendations of trust metric model
As previously mentioned, according to scoring and the viewing behavior of each user, user can be obtained to the explicit of a film or implicit expression marking, thus build the user-project matrix in window seclected time.In general, user-project matrix is sparse matrix, and in order to solve the openness problem of user-project matrix, the present invention uses random walk model to replace collaborative filtering method to carry out individualized video recommendation.
1) Random Walker algorithm
Random Walker algorithm uses Markov model, migration from a point, until travel through complete set.At any one point, all can with probability 1-Pr (X
u, k+1| X
u,k) migration to certain neighbor node, with probability P r (X
u, k+1| X
u,k) random skip to set in any point (comprising self).Often carry out a migration, arrive the probability of certain node only by current place Determines.Wherein, Pr (X
u, k+1| X
u,k) be the transition probability parameter of random walk, X
u, k+1represent in u situation, the state of k+1 time point; X
u,krepresent in u situation, the state of k time point.
The migration each time of random walk can obtain a probability distribution, and this distribution features the probability that in set, each point is accessed to.The process of whole random walk is exactly that when meeting certain condition, this distribution probability can be tending towards convergence by above-mentioned probability distribution as the input of next migration and the process iterated, thus obtains a stable probability distribution.Random walk model is used widely in fields such as data minings, and such as classical PageRank algorithm is exactly an example of random walk.
2) individualized video based on random walk model is recommended
Suppose that I represents the Item Sets of video presentation in individualized video commending system, m is the size of Item Sets, then can obtain the oriented weighted graph of m node, way interior joint node
irepresent project i, limit edge
i,jweight P
i,jthe probability P of project j is transferred in representative from project i
i,j=Pr (X
u, k+1=j|X
u,k=i)
According to above-mentioned transition probability, the item advance probability matrix P on m*m rank can be obtained, the transition probability vector of each row representative from m item advance to project j in matrix.If U represents user's collection of commending system, n is the scale that user collects, and so can set up the initial user-project rating matrix R on n*m rank according to the history scoring record of user, wherein the scoring vector of every a line representative of consumer, each element R
uirepresentative of consumer u, to the history score data of project i, sets R
ui=Pr (X
u, 0=i).Then can calculate user u according to user-project rating matrix and walk the transition probability being in project j in kth:
In above formula, the α factor is the probability that user proceeds next step under current state, and this parameter can control the step-length of random walk, and rule of thumb, this numerical value should increase along with the reduction of training set ratio.The overall probability of user u to project j can be obtained thus:
Wherein c represents constant, can calculate the overall probability of all users to all items, thus can obtain user-entry sorting matrix according to above formula
By means of this matrix can estimating user to the interest level of any one project, thus for individualized video recommend reference frame is provided.
Above embodiment is only in order to illustrate technical scheme of the present invention but not to be limited; those of ordinary skill in the art can modify to technical scheme of the present invention or equivalent replacement; and not departing from the spirit and scope of the present invention, protection scope of the present invention should be as the criterion with described in claims.
Claims (9)
1., based on a video recommendation method for timing driving, its step comprises:
1) by the interest graded of third party's data analysis user to video, and user interest gradient curve is obtained, using the time point that the singular point of user interest gradient curve moves as user interest;
2) determine the interest transit time point that user is nearest, the user-project scoring later to the interest transit time point that user is nearest gathers, thus sets up the user the meeting user's current interest-project rating matrix in window seclected time;
3) based on described user-project rating matrix, the individualized video using random walk model to carry out user is recommended.
2. the method for claim 1, is characterized in that: step 1) described third party's data are bean cotyledon scoring, draw user interest gradient curve by the bean cotyledon scoring history analyzing user.
3. method as claimed in claim 1 or 2, is characterized in that, step 1) method that obtains described user interest gradient curve is: each attribute of video is formed an attribute vector; At each time point, each attribute of video is weighted according to current time and the user of time before record of marking, obtains a user interest vector; According to a series of user interests vector of user's multiple time point on a timeline, analyze and obtain time dependent user interest gradient curve.
4. method as claimed in claim 3, it is characterized in that, the computing formula of the interest gradient of certain time point t is:
D(t)=I(t)-I(t-1),
Wherein, D (t) the user interest gradient that is t; The user interest vector that I (t) is t; The user interest vector that I (t-1) is t previous moment.
5. method as claimed in claim 3, it is characterized in that, each attribute of described video comprises: title, performer, director, age, brief introduction.
6. method as claimed in claim 1 or 2, is characterized in that, step 2) described user-project scoring comprises explicit scoring and implicit scores; Described explicit scoring is when user watches video, the scoring that user carries out the current video watched; Described implicit scores be recording user viewing, redirect, collect, share action, according to the different action prediction users of user to the fancy grade of video.
7. method as claimed in claim 6, it is characterized in that, step 2) by quantizing and normalization process user-project scoring, obtain interest transit time later user-project rating matrix, every a line in matrix represents a user characteristics, and each row represents an item characteristic.
8. method as claimed in claim 1 or 2, it is characterized in that, step 3) first calculate user u according to user-project rating matrix and be in the transition probability of project j in kth step, then user u is calculated to the overall probability of project j, then calculate the overall probability of all users to all items, thus obtain user-entry sorting matrix
and carry out individualized video recommendation according to this matrix.
9. method as claimed in claim 8, is characterized in that, step 3) concrete grammar be:
Suppose that I represents the Item Sets of video presentation in individualized video commending system, m is the size of Item Sets, first obtains the oriented weighted graph of m node, way interior joint node
irepresent project i, limit edge
i,jweight P
i,jthe probability P of project j is transferred in representative from project i
i,j=Pr (X
u, k+1=j|X
u,k=i);
According to above-mentioned transition probability, obtain the item advance probability matrix P on m*m rank, the transition probability vector of each row representative from m item advance to project j in matrix; If U represents user's collection of commending system, n is the scale that user collects, and sets up the initial user-project rating matrix R on n*m rank, wherein the scoring vector of every a line representative of consumer, each element R according to the history scoring record of user
uirepresentative of consumer u, to the history score data of project i, sets R
ui=Pr (X
u, 0=i), calculate user u according to user-project rating matrix and walk the transition probability being in project j in kth:
Wherein, the α factor is the probability that user proceeds next step under current state; The overall probability of user u to project j can be obtained thus:
Wherein c represents constant, calculates the overall probability of all users to all items, thus obtain user-entry sorting matrix according to above formula
According to the interest level of this Matrix Estimation user to any one project, thus the individualized video realizing user is recommended.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510290170.2A CN104935963B (en) | 2015-05-29 | 2015-05-29 | A kind of video recommendation method based on timing driving |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510290170.2A CN104935963B (en) | 2015-05-29 | 2015-05-29 | A kind of video recommendation method based on timing driving |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104935963A true CN104935963A (en) | 2015-09-23 |
CN104935963B CN104935963B (en) | 2018-03-16 |
Family
ID=54122874
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510290170.2A Active CN104935963B (en) | 2015-05-29 | 2015-05-29 | A kind of video recommendation method based on timing driving |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104935963B (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105447087A (en) * | 2015-11-06 | 2016-03-30 | 腾讯科技(深圳)有限公司 | Video recommendation method and apparatus |
CN105740473A (en) * | 2016-03-14 | 2016-07-06 | 腾讯科技(深圳)有限公司 | User-generated content display method and device |
CN105913296A (en) * | 2016-04-01 | 2016-08-31 | 北京理工大学 | Customized recommendation method based on graphs |
CN106168980A (en) * | 2016-07-26 | 2016-11-30 | 合网络技术(北京)有限公司 | Multimedia resource recommends sort method and device |
CN106250499A (en) * | 2016-08-02 | 2016-12-21 | 合网络技术(北京)有限公司 | A kind of video is to method for digging and device |
CN106372101A (en) * | 2016-07-20 | 2017-02-01 | Tcl集团股份有限公司 | Video recommendation method and apparatus |
WO2017107802A1 (en) * | 2015-12-24 | 2017-06-29 | 阿里巴巴集团控股有限公司 | Method and device for associating network item and calculating association information |
CN107220329A (en) * | 2017-05-24 | 2017-09-29 | 苏州百智通信息技术有限公司 | The evaluation method and device of video resource |
WO2017219812A1 (en) * | 2016-06-25 | 2017-12-28 | 华为技术有限公司 | Content recommendation method and device |
CN108170868A (en) * | 2018-02-09 | 2018-06-15 | 宁夏灵智科技有限公司 | Video recommendation method and device |
CN108257035A (en) * | 2017-12-30 | 2018-07-06 | 黄敬易 | User's similarity calculating method, device, server and readable storage medium storing program for executing |
CN108415987A (en) * | 2018-02-12 | 2018-08-17 | 大连理工大学 | A kind of cold start-up solution that film is recommended |
CN108628999A (en) * | 2018-05-02 | 2018-10-09 | 南京大学 | It is a kind of based on explicit and implicit information video recommendation method |
CN109189988A (en) * | 2018-09-18 | 2019-01-11 | 北京邮电大学 | A kind of video recommendation method |
CN109389168A (en) * | 2018-09-29 | 2019-02-26 | 国信优易数据有限公司 | Project recommendation model training method, item recommendation method and device |
CN109587527A (en) * | 2018-11-09 | 2019-04-05 | 青岛聚看云科技有限公司 | A kind of method and device that individualized video is recommended |
CN109800328A (en) * | 2019-01-08 | 2019-05-24 | 青岛聚看云科技有限公司 | Video recommendation method, its device, information processing equipment and storage medium |
CN110175286A (en) * | 2019-05-17 | 2019-08-27 | 山东师范大学 | It is combined into the Products Show method and system to optimization and matrix decomposition |
CN110390415A (en) * | 2018-04-18 | 2019-10-29 | 北京嘀嘀无限科技发展有限公司 | A kind of method and system carrying out trip mode recommendation based on user's trip big data |
CN111292151A (en) * | 2018-12-07 | 2020-06-16 | 中移互联网有限公司 | Object recommendation method and device, storage medium and object recommendation equipment |
CN109543111B (en) * | 2018-11-28 | 2021-09-21 | 广州虎牙信息科技有限公司 | Recommendation information screening method and device, storage medium and server |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101551825A (en) * | 2009-05-15 | 2009-10-07 | 中国科学技术大学 | Personalized film recommendation system and method based on attribute description |
WO2013046220A2 (en) * | 2011-06-28 | 2013-04-04 | Tata Consultancy Services Limited | A method and system for context-aware recommendation |
CN104063481A (en) * | 2014-07-02 | 2014-09-24 | 山东大学 | Film individuation recommendation method based on user real-time interest vectors |
-
2015
- 2015-05-29 CN CN201510290170.2A patent/CN104935963B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101551825A (en) * | 2009-05-15 | 2009-10-07 | 中国科学技术大学 | Personalized film recommendation system and method based on attribute description |
WO2013046220A2 (en) * | 2011-06-28 | 2013-04-04 | Tata Consultancy Services Limited | A method and system for context-aware recommendation |
CN104063481A (en) * | 2014-07-02 | 2014-09-24 | 山东大学 | Film individuation recommendation method based on user real-time interest vectors |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105447087A (en) * | 2015-11-06 | 2016-03-30 | 腾讯科技(深圳)有限公司 | Video recommendation method and apparatus |
CN105447087B (en) * | 2015-11-06 | 2021-08-24 | 腾讯科技(深圳)有限公司 | Video recommendation method and device |
WO2017107802A1 (en) * | 2015-12-24 | 2017-06-29 | 阿里巴巴集团控股有限公司 | Method and device for associating network item and calculating association information |
CN105740473A (en) * | 2016-03-14 | 2016-07-06 | 腾讯科技(深圳)有限公司 | User-generated content display method and device |
CN105913296A (en) * | 2016-04-01 | 2016-08-31 | 北京理工大学 | Customized recommendation method based on graphs |
WO2017219812A1 (en) * | 2016-06-25 | 2017-12-28 | 华为技术有限公司 | Content recommendation method and device |
CN106372101A (en) * | 2016-07-20 | 2017-02-01 | Tcl集团股份有限公司 | Video recommendation method and apparatus |
CN106372101B (en) * | 2016-07-20 | 2019-09-17 | Tcl集团股份有限公司 | A kind of video recommendation method and device |
CN106168980A (en) * | 2016-07-26 | 2016-11-30 | 合网络技术(北京)有限公司 | Multimedia resource recommends sort method and device |
CN106168980B (en) * | 2016-07-26 | 2020-07-28 | 阿里巴巴(中国)有限公司 | Multimedia resource recommendation sequencing method and device |
CN106250499A (en) * | 2016-08-02 | 2016-12-21 | 合网络技术(北京)有限公司 | A kind of video is to method for digging and device |
CN107220329A (en) * | 2017-05-24 | 2017-09-29 | 苏州百智通信息技术有限公司 | The evaluation method and device of video resource |
CN107220329B (en) * | 2017-05-24 | 2021-01-19 | 苏州百智通信息技术有限公司 | Video resource evaluation method and device |
CN108257035A (en) * | 2017-12-30 | 2018-07-06 | 黄敬易 | User's similarity calculating method, device, server and readable storage medium storing program for executing |
CN108170868A (en) * | 2018-02-09 | 2018-06-15 | 宁夏灵智科技有限公司 | Video recommendation method and device |
CN108415987B (en) * | 2018-02-12 | 2021-05-11 | 大连理工大学 | Cold start solution for movie recommendation |
CN108415987A (en) * | 2018-02-12 | 2018-08-17 | 大连理工大学 | A kind of cold start-up solution that film is recommended |
CN110390415A (en) * | 2018-04-18 | 2019-10-29 | 北京嘀嘀无限科技发展有限公司 | A kind of method and system carrying out trip mode recommendation based on user's trip big data |
US11151680B2 (en) | 2018-04-18 | 2021-10-19 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for recommending transportation means |
CN108628999B (en) * | 2018-05-02 | 2022-11-11 | 南京大学 | Video recommendation method based on explicit and implicit information |
CN108628999A (en) * | 2018-05-02 | 2018-10-09 | 南京大学 | It is a kind of based on explicit and implicit information video recommendation method |
CN109189988B (en) * | 2018-09-18 | 2021-06-22 | 北京邮电大学 | Video recommendation method |
CN109189988A (en) * | 2018-09-18 | 2019-01-11 | 北京邮电大学 | A kind of video recommendation method |
CN109389168A (en) * | 2018-09-29 | 2019-02-26 | 国信优易数据有限公司 | Project recommendation model training method, item recommendation method and device |
CN109587527A (en) * | 2018-11-09 | 2019-04-05 | 青岛聚看云科技有限公司 | A kind of method and device that individualized video is recommended |
CN109587527B (en) * | 2018-11-09 | 2021-06-04 | 青岛聚看云科技有限公司 | Personalized video recommendation method and device |
CN109543111B (en) * | 2018-11-28 | 2021-09-21 | 广州虎牙信息科技有限公司 | Recommendation information screening method and device, storage medium and server |
CN111292151A (en) * | 2018-12-07 | 2020-06-16 | 中移互联网有限公司 | Object recommendation method and device, storage medium and object recommendation equipment |
CN111292151B (en) * | 2018-12-07 | 2023-07-25 | 中移互联网有限公司 | Object recommendation method and device, storage medium and object recommendation equipment |
CN109800328A (en) * | 2019-01-08 | 2019-05-24 | 青岛聚看云科技有限公司 | Video recommendation method, its device, information processing equipment and storage medium |
CN109800328B (en) * | 2019-01-08 | 2021-08-20 | 青岛聚看云科技有限公司 | Video recommendation method, device thereof, information processing equipment and storage medium |
CN110175286B (en) * | 2019-05-17 | 2021-05-11 | 山东师范大学 | Product recommendation method and system combining pairwise optimization and matrix decomposition |
CN110175286A (en) * | 2019-05-17 | 2019-08-27 | 山东师范大学 | It is combined into the Products Show method and system to optimization and matrix decomposition |
Also Published As
Publication number | Publication date |
---|---|
CN104935963B (en) | 2018-03-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104935963B (en) | A kind of video recommendation method based on timing driving | |
CN108256093B (en) | Collaborative filtering recommendation algorithm based on multiple interests and interest changes of users | |
Luo et al. | Personalized recommendation by matrix co-factorization with tags and time information | |
US9171078B2 (en) | Automatic recommendation of vertical search engines | |
KR101871747B1 (en) | Similarity tendency based user-sightseeing recommendation system and method thereof | |
TWI636416B (en) | Method and system for multi-phase ranking for content personalization | |
CN103533390B (en) | The method and system of television program recommendations are carried out based on social network information | |
CN110879864B (en) | Context recommendation method based on graph neural network and attention mechanism | |
CN104063481A (en) | Film individuation recommendation method based on user real-time interest vectors | |
CN103955535A (en) | Individualized recommending method and system based on element path | |
CN104391849A (en) | Collaborative filtering recommendation method for integrating time contextual information | |
KR20150023432A (en) | Method and apparatus for inferring user demographics | |
CN104899246A (en) | Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism | |
CN102156747B (en) | Method and device for forecasting collaborative filtering mark by introduction of social tag | |
CN105430505A (en) | IPTV program recommending method based on combined strategy | |
CN113420221B (en) | Interpretable recommendation method integrating implicit article preference and explicit feature preference of user | |
CN115048586B (en) | Multi-feature-fused news recommendation method and system | |
CN105761154A (en) | Socialized recommendation method and device | |
CN107545075A (en) | A kind of restaurant recommendation method based on online comment and context aware | |
CN113407729B (en) | Judicial-oriented personalized case recommendation method and system | |
Yu et al. | Point-of-interest recommendation based on user contextual behavior semantics | |
CN112364245B (en) | Top-K movie recommendation method based on heterogeneous information network embedding | |
Baby et al. | A hybrid product recommendation system based on weather analysis | |
CN105654345A (en) | Bayesian network-based advertisement click-through rate prediction method and device | |
CN114022233A (en) | Novel commodity recommendation method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |