CN107343209A - A kind of UGC associated video generation methods based on increment collaborative filtering - Google Patents
A kind of UGC associated video generation methods based on increment collaborative filtering Download PDFInfo
- Publication number
- CN107343209A CN107343209A CN201710517911.5A CN201710517911A CN107343209A CN 107343209 A CN107343209 A CN 107343209A CN 201710517911 A CN201710517911 A CN 201710517911A CN 107343209 A CN107343209 A CN 107343209A
- Authority
- CN
- China
- Prior art keywords
- videoid
- data
- video
- user
- count
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Graphics (AREA)
- Computing Systems (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a kind of UGC associated video generation methods based on increment collaborative filtering, including:Original user is read in real time and plays behavioral data, and is pre-processed, and exports user's played data;Calculated according to user's played data and export the final popularity data of video;Incremental data and increment co-occurrence data therein are obtained according to user's played data;Final co-occurrence data is calculated according to history co-occurrence data and increment co-occurrence data;For each video data pair, converging operation is carried out to two video datas of video data centering respectively, obtains last associated video list.Method in the present invention, user behavior data can be handled in real time, collaborative filtering recommending result can quickly be fed back to recommendation results and concentrated;The collaborative filtering result of new video is quickly calculated, the problem of avoiding the recommendation cold starting effect of new video bad.
Description
Technical field
The present invention relates to user behavior analysis and recommend method, more particularly to a kind of UGC phases based on increment collaborative filtering
Close video generation method.
Background technology
There is the video that user largely persistently uploads in video UGC websites, compared with non-UGC videos, UGC videos have duration
Shorter, the features such as renewal time is fast, life cycle is short.The associated recommendation of UGC videos refers to, according to a given video, recommend
The TOP K videos of (related) most like to its.The application scenarios of associated video result include:
1) personalized recommendation
User have viewed video A, recommend the results for video related to video A for it.
2) associated recommendation
The associated video that video A associated video result is placed in broadcast page recommends position.
Collaborative filtering is according to behavior of the user to article, the interest distribution of co-occurrence data and user between more new article
Data, to calculate the similarity between article between user, carry out the Collaborative Recommendation based on article or user.Associated video result can
Can be calculated using the collaborative filtering (Item Base CF, abbreviation ICF) based on article, that is, utilize the co-occurrence between video
Degree calculates the similarity of video between any two, so as to calculating maximally related K video.
Because ICF algorithms need substantial amounts of user behavior data to carry out off-line calculation, and the life cycle of UGC videos compared with
It is short, after new video i generations, as user constantly produces viewing behavior to it, it can be calculated and regarded by co-user viewing number
Frequency i associated recommendation results for video.Then due to the off-line calculation feature of traditional IC F algorithms, such as calculate once daily, due to being
Regularly update, therefore associated video result has serious hysteresis quality, causes associated video result empty window phase, while user to be present
Personalized recommendation result there is also it is not accurate the problem of.Therefore a kind of ICF methods of increment are needed, quickly, are caught in real time
User behavior, improve the ageing of ICF associated video results.
UGC video deltas ICF difficult point:
Constantly produced because user watches behavior, if all going to update associated video ICF results to each user behavior,
Substantial amounts of amount of calculation will be brought.Therefore need to select effective candidate's similar video data pair, for candidate video data pair
Their similitude is calculated, while needs to feed back to caused user behavior in ICF associated video results as soon as possible again.
The content of the invention
The technical problem to be solved in the present invention is to realize mountain-climbing record and the time-consuming statistics of mountain-climbing, bag by gps coordinate position
Include mountain-climbing point entry position, the mountain top title on every mountain;Recorded, stepped on come the mountain-climbing of counting user by APP and wechat public number
Mountain takes.
Solves above-mentioned technical problem, the invention provides a kind of UGC associated videos generation side based on increment collaborative filtering
Method, including:
Original user is read in real time and plays behavioral data UserPlayLog, and is pre-processed, and output user plays number
According to UserPlayData, UserPlayData forms are:Userid, [(videoid_1, ref_1), (videoid_2, ref_
2) ... (videoid_i, ref_i)], userid is user data, and videoid_i is the video data that the user watched,
Ref_i is user's score data with video data videoid_i;According to the userid in UserPlayData to watching behavior
It is polymerize, while the excessive user of viewing video number is filtered;
User played data UserPlayData is read, video data videoid_i in user's played data is gathered
Close, count the corresponding user's score data summation of each video data:Incremental video popularity and historical data are carried out
Merge, calculate and export the final popularity data VideoPopularityData of video;
User played data UserPlayData is read, obtaining viewing in all users according to user's played data appoints
The video data of two videos of meaning obtains to (videoid_i, videoid_j), and according to the video data to polymerizeing
RtCooccurData, field include (videoid_i, videoid_j), and rt_count_ij, wherein rt_count_ij are increment
Co-occurrence data, represent to watch the number of video videoid_i and videoid_j user simultaneously;
Incremental data RtCooccurData and historical data HisCooccurData is read, according to history co-occurrence data
His_count_ij and increment co-occurrence data rt_count_ij calculate final co-occurrence data count_ij, during calculating and to going through
The history data HisCooccurData fields that decay are (videoid_i, videoid_j), his_count_ij;
For each video data pair, respectively two video data videoid_i to video data centering and
Videoid_j carries out converging operation, obtains last associated video list.
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, it is described to wherein regarding
Frequency is polymerize according to videoid_i, is counted the corresponding user's score data summation of each video data and is specially:
Videoid ref summations are counted by following formula:
Wherein ref_i_new represents the total score of video i in incremental data, ref_kiRepresent k couples of user in incremental data
In video i score.
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, it is described to regard increment
Frequency popularity and historical data merge, and calculate and export the final popularity data VideoPopularityData of video
Specific method is:
For video data videoid_i, corresponding user's score data computational methods are as follows:
Ref_i=ref_i_new+a*ref_i_old
Wherein ref_i_old is video i history popularity, and a is decay factor, and 0<a<1;The popularity data
VideoPopularityData field includes:Videoid, ref.
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, the video data
It is to (videoid_i, videoid_j) building method:
Any two the video videoid_i and videoid_j that each user is watched in list of videos, it is configured in order
Video data is to (videoid_i, videoid_j).
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, the count_ij
Computational methods be shown below:
Wherein kijWhether expression user k watches video i and is video j, is then in this way 1, is otherwise 0.
Further, the UGC associated video generation methods based on increment collaborative filtering, the calculating are final as the aforementioned
Co-occurrence data count_ij specifically uses equation below:
Count_ij=a*his_count_ij+rt_count_ij
Wherein a is decay factor.
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, it is described to video counts
Carrying out converging operation according to two video datas videoid_i and videoid_j of centering is specially:
For each video data to (videoid_i, videoid_j), respectively according to videoid_i and videoid_
J carries out converging operation, and list is obtained for video data videoid_i:
[(videoid_1,count_1i),(videoid_2,count_2i),…..(videoid_n,count_ni)]
For video data videoid_i, the candidate video videoid_j following institute of correlation rel_ij calculations
Show:
Candidate's associated video is ranked up according to rel_ij, exports last associated video list
VideoRelateData, field include videoid_k, [(videoid_1, rel_k1), (videoid_2, ref_k2) ...
(videoid_n, ref_kn)], wherein ref_k1>=ref_k2>=... ref_kn.
Further, UGC associated video generation methods based on increment collaborative filtering as the aforementioned, in video data
Videoid_i obtains list:
[(videoid_1,count_1i),(videoid_2,count_2i),…..(videoid_n,count_ni)]
Also include afterwards:
Confidence level count_confidence is set to co-occurrence number, the video data that the value is less than for co-occurrence number enters
Row filtering, for reducing the list length.
Beneficial effects of the present invention:
1) the inventive method can handle user behavior data in real time, can quickly feed back collaborative filtering recommending result
To recommendation results and concentrate;
2) this method can quickly calculate the collaborative filtering result of new video, avoid the recommendation cold start-up of new video from imitating
The problem of fruit is bad;
3) this method avoid the complex operations of user's set of the viewing video of preservation, while also to the stream of recommendation video
Row degree carries out drop power to overheat video, and " hamlet " problem is avoided indirectly with smaller cost.
Brief description of the drawings
The step of Fig. 1 is the UGC associated video generation methods based on increment collaborative filtering in an embodiment of the present invention
Schematic diagram;
Fig. 2 is the detailed of the UGC associated video generation methods based on increment collaborative filtering in an embodiment of the present invention
Schematic flow sheet;
Fig. 3 is a kind of schematic diagram of embodiment of step S102 in Fig. 1;
Fig. 4 is a kind of schematic diagram of embodiment of step S103 in Fig. 1;
Fig. 5 is a kind of schematic diagram of embodiment of step S104 in Fig. 1;
Fig. 6 is a kind of schematic diagram of embodiment of step S105 in Fig. 1.
Embodiment
For the object, technical solutions and advantages of the present invention are more clearly understood, below in conjunction with specific embodiment, and reference
Accompanying drawing, the present invention is described in more detail.
Fig. 2 is that the flow of the UGC associated video generation methods based on increment collaborative filtering in one embodiment of the invention is shown
It is intended to.
As shown in Fig. 1 step schematic diagrams, a kind of UGC associated video generation methods based on increment collaborative filtering, including:
S101, original user broadcasting behavioral data UserPlayLog is read in real time, and pre-processed, export user
Played data UserPlayData, UserPlayData form is:Userid, [(videoid_1, ref_1), (videoid_2,
Ref_2) ... (videoid_i, ref_i)], userid is user data, and videoid_i is the video counts that the user watched
According to ref_i is user's score data with video data videoid_i;Viewing is gone according to the userid of original incremental data
To be polymerize, while the excessive user of viewing video number is filtered;
S102, user played data UserPlayData is read, video data videoid_i in user's played data is entered
Row polymerization, counts the corresponding user's score data summation of each video data:By incremental video popularity and history video
Popularity merges, and calculates and exports the final popularity data VideoPopularityData of video;
S103, user played data UserPlayData is read, obtained according to user's played data in all users
The video data of any two video is watched to (videoid_i, videoid_j), and according to the video data to gathering
Close, obtain RtCooccurData, field includes (videoid_i, videoid_j), rt_count_ij, wherein rt_count_
Ij is increment co-occurrence data, represents to watch the number of video videoid_i and videoid_j user simultaneously;
S104, incremental data RtCooccurData and historical data HisCooccurData is read, according to history co-occurrence number
Calculate final co-occurrence data count_ij according to his_count_ij and increment co-occurrence data rt_count_ij, it is during calculating and right
The historical data HisCooccurData fields that decay are (videoid_i, videoid_j), his_count_ij;
S105, for each video data pair, respectively two video data videoid_i to video data centering and
Videoid_j carries out converging operation, obtains last associated video list.
As shown in figure 3, a kind of above-mentioned steps S102 concrete methods of realizing is as described below:
S201, read user's played data UserPlayData;
S202, described wherein video data videoid_i is polymerize;
S203, to count by following formula the corresponding user's score data ref of each video data videoid_i total
With:
Wherein ref_i_new represents the total score of video i in incremental data, and ref_ki represents k couples of user in incremental data
In video i score;
S204, incremental video popularity and historical video streams row degree merged, calculate and export the final stream of video
Row degrees of data VideoPopularityData, specific method are:
For video data videoid_i, corresponding user's score data computational methods are as follows:
Ref_i=ref_i_new+a*ref_i_old
Wherein ref_i_old is video i history popularity, and a is decay factor, and 0<a<1;The popularity data
VideoPopularityData field includes:Videoid, ref.
As shown in figure 4, a kind of above-mentioned steps S103 concrete methods of realizing is as described below:
S301, read user's played data UserPlayData;
S302, any two the video videoid_i and videoid_j for watching each user in list of videos, construction
Into orderly video data to (videoid_i, videoid_j), wherein videoid_i<videoid_j;
S303, according to the video data (videoid_i, videoid_j) is polymerize, obtain (videoid_i,
Videoid_j), count_ij, wherein count_ij are co-occurrence data, represent simultaneously watched video videoid_i and
The number of videoid_j user;The computational methods of the count_ij are shown below:
Wherein kijWhether expression user k watches video i and is video j, is then in this way 1, is otherwise 0;
S304, output increment data RtCooccurData, field include (videoid_i, videoid_j), count_
ij。
As shown in figure 5, a kind of above-mentioned steps S104 concrete methods of realizing is as described below:
S401, incremental data RtCooccurData and historical data HisCooccurData is read, it is described
HisCooccurData fields are (videoid_i, videoid_j), his_count_ij
S402, historical data is decayed;
S403, final co-occurrence data count_ij, specific use are calculated according to history co-occurrence data and increment co-occurrence data
Equation below:
Count_ij=a*his_count_ij+rt_count_ij
Wherein a is decay factor.
As shown in fig. 6, a kind of above-mentioned steps S105 concrete methods of realizing is as described below:
S501, for each video data to (videoid_i, videoid_j), respectively according to videoid_i and
Videoid_j carries out converging operation, and list is obtained for video data videoid_i;
[(videoid_1,count_1i),(videoid_2,count_2i),…..(videoid_n,count_ni)];
S502, in order to reduce list length, can to co-occurrence number set confidence level count_confidence, for altogether
The video that occurrence number is less than the value is filtered;
S503, for video data videoid_i, candidate video videoid_j correlation rel_ij calculations are such as
Shown in lower:
The popularity being used here in VideoPopularityData, popularity drop power is carried out to candidate video;
S504, according to rel_ij candidate's associated video is ranked up, exports last associated video list
VideoRelateData, field include videoid_k, [(videoid_1, rel_k1), (videoid_2, ref_k2) ...
(videoid_n, ref_kn)], wherein ref_k1>=ref_k2>=... ref_kn, it is preferred that and n length is generally 102Rule
Mould.
Those of ordinary skills in the art should understand that:More than, described is only the specific embodiment of the present invention, and
The limitation present invention is not used in, within the spirit and principles of the invention, any modification, equivalent substitution and improvements done etc.,
It should be included within protection scope of the present invention.The inventive method or equivalent technical principle all belong within protection domain.
Claims (8)
- A kind of 1. UGC associated video generation methods based on increment collaborative filtering, it is characterised in that including:Original user is read in real time and plays behavioral data UserPlayLog, and is pre-processed, and exports user's played data UserPlayData, UserPlayData form is:Userid, [(videoid_1, ref_1), (videoid_2, ref_ 2) ... (videoid_i, ref_i)], userid is user data, and videoid_i is the video data that the user watched, Ref_i is user's score data with video data videoid_i;According in user's played data UserPlayData Userid polymerize to viewing behavior, while the excessive user of viewing video number is filtered;User played data UserPlayData is read, video data videoid_i in user's played data is polymerize, is united Count out the corresponding user's score data summation of each video data:By incremental video popularity ref_i_new and history video Popularity ref_i_old is merged, and is calculated and is exported the final popularity data VideoPopularityData of video;User played data UserPlayData is read, is obtained according to user's played data in all users and watches any two The video data of individual video obtains incremental data to (videoid_i, videoid_j), and according to video data to polymerizeing RtCooccurData, field include (videoid_i, videoid_j), and rt_count_ij, wherein rt_count_ij are increment Co-occurrence data, represent to watch the number of video videoid_i and videoid_j user simultaneously;Incremental data RtCooccurData and historical data HisCooccurData is read, according to history co-occurrence data his_ Count_ij and increment co-occurrence data rt_count_ij calculate final co-occurrence data count_ij, during calculating and to historical data Decayed;The HisCooccurData fields are (videoid_i, videoid_j), his_count_ij;For each video data pair, respectively to two video datas videoid_i and videoid_j of video data centering Converging operation is carried out, obtains last associated video list.
- 2. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute State and wherein video data videoid_i is polymerize, count the corresponding user's score data summation of each video data Specially:Videoid ref summations are counted by following formula:<mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> <mo>_</mo> <mi>i</mi> <mo>_</mo> <mi>n</mi> <mi>e</mi> <mi>w</mi> <mo>=</mo> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </msubsup> <mi>r</mi> <mi>e</mi> <mi>f</mi> <mo>_</mo> <mi>k</mi> <mi>i</mi> <mo>,</mo> <mrow> <mo>(</mo> <mi>r</mi> <mi>e</mi> <mi>f</mi> <mo>_</mo> <mi>k</mi> <mi>i</mi> <mo>></mo> <mo>=</mo> <mn>0</mn> <mo>)</mo> </mrow> </mrow>Wherein ref_i_new represent incremental data in video i total score, ref_ki represent incremental data in user k for regarding Frequency i score.
- 3. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute State and merge incremental video popularity and historical data, calculate and export the final popularity data of video VideoPopularityData specific methods are:For video data videoid_i, corresponding user's score data computational methods are as follows:Ref_i=ref_i_new+a*ref_i_oldWherein ref_i_old is video i history popularity, and a is decay factor, and 0<a<1;The popularity data VideoPopularityData field includes:Videoid, ref.
- 4. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute State video data is to (videoid_i, videoid_j) building method:Any two the video videoid_i and videoid_j that each user is watched in list of videos, are configured to orderly video Data are to (videoid_i, videoid_j).
- 5. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute The computational methods for stating rt_count_ij are shown below:<mrow> <mi>r</mi> <mi>t</mi> <mo>_</mo> <mi>c</mi> <mi>o</mi> <mi>u</mi> <mi>n</mi> <mi>t</mi> <mo>_</mo> <mi>i</mi> <mi>j</mi> <mo>=</mo> <munderover> <mo>&Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>m</mi> </munderover> <msub> <mi>k</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> </mrow>Wherein kijWhether expression user k watches video i and is video j, is then in this way 1, is otherwise 0.
- 6. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute State the final co-occurrence data count_ij of calculating and specifically use equation below:Count_ij=a*his_count_ij+rt_count_ijWherein a is decay factor.
- 7. the UGC associated video generation methods according to claim 1 based on increment collaborative filtering, it is characterised in that institute State and be specially to two video datas videoid_i and videoid_j the progress converging operation of video data centering:For each video data to (videoid_i, videoid_j), enter respectively according to videoid_i and videoid_j Row converging operation, list is obtained for video data videoid_i:[(videoid_1,count_1i),(videoid_2,count_2i),…..(videoid_n,count_ni)]For video data videoid_i, candidate video videoid_j correlation rel_ij calculations are as follows:<mrow> <mi>r</mi> <mi>e</mi> <mi>l</mi> <mo>_</mo> <mi>i</mi> <mi>j</mi> <mo>=</mo> <mfrac> <mrow> <mi>c</mi> <mi>o</mi> <mi>u</mi> <mi>n</mi> <mi>t</mi> <mo>_</mo> <mi>i</mi> <mi>j</mi> </mrow> <mrow> <mi>r</mi> <mi>e</mi> <mi>f</mi> <mo>_</mo> <mi>j</mi> </mrow> </mfrac> </mrow>Candidate's associated video is ranked up according to rel_ij, exports last associated video list VideoRelateData, word Section includes videoid_k, [(videoid_1, rel_k1), (videoid_2, ref_k2) ... (videoid_n, ref_kn)], Wherein ref_k1>=ref_k2>=... ref_kn.
- 8. the UGC associated video generation methods according to claim 7 based on increment collaborative filtering, it is characterised in that Video data videoid_i obtains list:[(videoid_1, count_1i), (videoid_2, count_2i) ... .. (videoid_n, count_ni)] go back afterwards Including:Confidence level count_confidence is set to co-occurrence number, the video data that the value is less than for co-occurrence number was carried out Filter, for reducing the list length.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710517911.5A CN107343209A (en) | 2017-06-29 | 2017-06-29 | A kind of UGC associated video generation methods based on increment collaborative filtering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710517911.5A CN107343209A (en) | 2017-06-29 | 2017-06-29 | A kind of UGC associated video generation methods based on increment collaborative filtering |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107343209A true CN107343209A (en) | 2017-11-10 |
Family
ID=60219267
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710517911.5A Pending CN107343209A (en) | 2017-06-29 | 2017-06-29 | A kind of UGC associated video generation methods based on increment collaborative filtering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107343209A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111400546A (en) * | 2020-03-18 | 2020-07-10 | 腾讯科技(深圳)有限公司 | Video recall method and video recommendation method and device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103209342A (en) * | 2013-04-01 | 2013-07-17 | 电子科技大学 | Collaborative filtering recommendation method introducing video popularity and user interest change |
CN104219575A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Related video recommending method and system |
CN104216884A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Collaborative filtering system and method on basis of time decay |
CN106570090A (en) * | 2016-10-20 | 2017-04-19 | 杭州电子科技大学 | Method for collaborative filtering recommendation based on interest changes and trust relations |
-
2017
- 2017-06-29 CN CN201710517911.5A patent/CN107343209A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103209342A (en) * | 2013-04-01 | 2013-07-17 | 电子科技大学 | Collaborative filtering recommendation method introducing video popularity and user interest change |
CN104219575A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Related video recommending method and system |
CN104216884A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津)科技有限公司 | Collaborative filtering system and method on basis of time decay |
CN106570090A (en) * | 2016-10-20 | 2017-04-19 | 杭州电子科技大学 | Method for collaborative filtering recommendation based on interest changes and trust relations |
Non-Patent Citations (1)
Title |
---|
黄学峰: "基于hadoop的电影推荐系统研究与实现", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111400546A (en) * | 2020-03-18 | 2020-07-10 | 腾讯科技(深圳)有限公司 | Video recall method and video recommendation method and device |
CN111400546B (en) * | 2020-03-18 | 2020-12-01 | 腾讯科技(深圳)有限公司 | Video recall method and video recommendation method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN100581227C (en) | Collaborative filtered recommendation method introducing hotness degree weight of program | |
CN103559206B (en) | A kind of information recommendation method and system | |
CN103678518B (en) | Method and device for adjusting recommendation lists | |
CN106561054B (en) | Recommend method and system in a kind of live streaming room for webcast website | |
CN101482884A (en) | Cooperation recommending system based on user predilection grade distribution | |
CN102630052B (en) | Real time streaming-oriented television program recommendation system | |
CN102207972B (en) | Television program recommending method and device for digital television | |
CN105975472A (en) | Method and device for recommendation | |
CN105430505B (en) | A kind of IPTV program commending methods based on combined strategy | |
CN103995839A (en) | Commodity recommendation optimizing method and system based on collaborative filtering | |
CN108134950B (en) | Intelligent video recommendation method and system | |
CN106846094A (en) | A kind of method and apparatus for recommending application message based on application has been installed | |
CN103491441A (en) | Recommendation method and system of live television programs | |
CN106096047B (en) | User partition preference calculation method and system based on Information Entropy | |
CN107748759A (en) | Video pushing method and its equipment | |
CN105721899A (en) | Video quality scoring method and system | |
CN103294812A (en) | Commodity recommendation method based on mixed model | |
CN105338408B (en) | Video recommendation method based on time factor | |
CN104298787A (en) | Individual recommendation method and device based on fusion strategy | |
CN107608990A (en) | A kind of live personalized recommendation method | |
CN106454431A (en) | Method and system for recommending television programs | |
CN106570031A (en) | Service object recommending method and device | |
CN109508407A (en) | The tv product recommended method of time of fusion and Interest Similarity | |
CN105373600A (en) | Method and device for sorting video playlists | |
CN108419134A (en) | The recommendation of the channels method merged with group current behavior based on individual history |
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 |
Application publication date: 20171110 |
|
RJ01 | Rejection of invention patent application after publication |