CN110351580B - Television program topic recommendation method and system based on non-negative matrix factorization - Google Patents
Television program topic recommendation method and system based on non-negative matrix factorization Download PDFInfo
- Publication number
- CN110351580B CN110351580B CN201910631011.2A CN201910631011A CN110351580B CN 110351580 B CN110351580 B CN 110351580B CN 201910631011 A CN201910631011 A CN 201910631011A CN 110351580 B CN110351580 B CN 110351580B
- Authority
- CN
- China
- Prior art keywords
- data
- label
- thematic
- matrix
- television program
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
- H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
- H04N21/235—Processing of additional data, e.g. scrambling of additional data or processing content descriptors
- H04N21/2355—Processing of additional data, e.g. scrambling of additional data or processing content descriptors involving reformatting operations of additional data, e.g. HTML pages
-
- 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
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/435—Processing of additional data, e.g. decrypting of additional data, reconstructing software from modules extracted from the transport stream
- H04N21/4355—Processing of additional data, e.g. decrypting of additional data, reconstructing software from modules extracted from the transport stream involving reformatting operations of additional data, e.g. HTML pages on a television screen
-
- 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
- H04N21/4508—Management of client data or end-user data
- H04N21/4532—Management of client data or end-user data involving end-user characteristics, e.g. viewer profile, 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
- H04N21/4668—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)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Computer Graphics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of big data, aims to solve the problem that the conventional television program recommendation cannot reasonably cover user portraits, and provides a television program topic recommendation method based on non-negative matrix decomposition, which comprises the following steps: the method comprises the steps of obtaining label information in real time when a television program is to be recommended to a user, wherein the label information comprises: user portrait data, program tag data, and thematic tag data, the user portrait data being used to represent a user's score for each tag; constructing a non-negative matrix decomposition model, converting label information into a full label matrix, and performing decomposition and dimension reduction processing on the full label matrix according to the non-negative matrix decomposition model to obtain a probability matrix; and reconstructing user portrait data and thematic label data according to the label information in the probability matrix, and recommending the television program thematic according to the reconstructed user portrait data and thematic label data. According to the invention, through the combination of the interest fields, the reasonable distribution of the user recommendation list in the interest fields is realized.
Description
Technical Field
The invention relates to the technical field of big data, in particular to a television program topic recommendation method and system.
Background
In personalized recommendation of the smart television, the problem that the labels intersect in the interest field is not noticed, namely, some labels are preferred by users with the same interest psychology, such as action classes and martial arts, and generally the action piece is scored high and the martial arts are scored high. However, in the algorithm for generating the recommendation list, the scores of the tags are mainly sorted according to the user, and if the scores of the user for swordsmen, action, antique and the like are high, the tags covered by the recommendation list programs are mostly the same. But there are also tags that have user scores centered on top, e.g., inspirations, adventures, comedies, etc. The proportion of the programs of the subjects such as the swordsmen action is too large, so that the recommendation space for the motivation, the adventure, the comedy and the like is extremely small, and the possibility of pushing the parts is extremely small. Therefore, the labels included in the recommended programs cannot reasonably cover the user portrait, and the recommendation effect is reduced.
non-Negative Matrix Factorization (NMF), proposed by Lee and Seung in the journal of nature 1999, is a Matrix factorization method that makes all the components after factorization non-negative (requiring purely additive description) and at the same time implements nonlinear dimensionality reduction. NMF has become one of the most popular multidimensional data processing tools in the research fields of signal processing, biomedical engineering, pattern recognition, computer vision, and image engineering.
Disclosure of Invention
The invention aims to solve the problem that the conventional television program recommendation cannot reasonably cover the portrait of a user, and provides a television program topic recommendation method and system based on non-negative matrix decomposition.
The technical scheme adopted by the invention for solving the technical problems is as follows: the television program topic recommendation method based on non-negative matrix factorization comprises the following steps:
step 1, obtaining label information in real time when a television program is to be recommended to a user, wherein the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information;
step 2, constructing a non-negative matrix decomposition model, converting the label information into a full label matrix, and performing decomposition and dimension reduction processing on the full label matrix according to the non-negative matrix decomposition model to obtain a probability matrix, wherein the probability matrix is label information generated by merging according to the relevance between thematic labels;
and 3, reconstructing user portrait data and thematic label data according to the label information in the probability matrix, and recommending the television program thematic according to the reconstructed user portrait data and thematic label data.
Further, in order to realize the decomposition of the full label matrix, in step 2, the decomposing and dimension reduction processing is performed on the full label matrix through the nonnegative matrix decomposition model to obtain the probability matrix specifically:
and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix subjected to full label matrix decomposition and dimension reduction.
Further, in order to implement recommendation of tv program topics, in step 3, the recommending of tv program topics according to the reconstructed user portrait data and topic label data includes:
calculating cosine similarity between reconstructed user portrait data and thematic label data, sorting the cosine similarity from big to small, and selecting and recommending television program thematic corresponding to N thematic label data before the cosine similarity ranking, wherein N is a positive integer greater than or equal to 1.
Further, to realize the calculation of the cosine similarity, the cosine similarity calculation formula is as follows:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
The invention also provides a television program topic recommendation system based on non-negative matrix factorization, which comprises the following steps:
the system comprises an acquisition unit and a recommendation unit, wherein the acquisition unit is used for acquiring label information in real time when a television program is to be recommended to a user, and the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information;
the nonnegative matrix decomposition model is used for performing decomposition and dimension reduction processing on a full label matrix to obtain a probability matrix, the full label matrix is obtained by converting the label information, and the probability matrix is label information generated by combining the label information according to the association degree between thematic labels;
and the recommending unit is used for reconstructing the user portrait data and the thematic label data according to the label information in the probability matrix and recommending the television program thematic according to the reconstructed user portrait data and the thematic label data.
Further, the decomposing and dimension reduction processing on the full label matrix to obtain the probability matrix specifically includes:
and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix after decomposition and dimension reduction processing is performed on the full label matrix.
Further, the recommending unit is further configured to:
calculating cosine similarity between reconstructed user portrait data and thematic label data, sorting the cosine similarity from big to small, and selecting and recommending television program thematic corresponding to N thematic label data before the cosine similarity ranking, wherein N is a positive integer greater than or equal to 1.
Further, the cosine similarity calculation formula is as follows:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
The invention has the beneficial effects that: according to the television program topic recommendation method and system based on non-negative matrix decomposition, a full label matrix is decomposed based on a non-negative matrix decomposition model, a probability matrix containing user portrait data and topic label data after dimension reduction is obtained, labels are optimized, and reasonable distribution of a user recommendation list in an interest field is finally achieved through combination of the interest field.
Drawings
FIG. 1 is a schematic diagram of a non-negative matrix factorization;
FIG. 2 is a schematic flow chart of a non-negative matrix factorization-based television program topic recommendation method according to the present invention;
fig. 3 is a schematic structural diagram of the television program topic recommendation system based on non-negative matrix factorization according to the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Various dimension reduction methods are contrastively analyzed, and in the dimension reduction of high-dimensional data, the distance-based dimension reduction method shows that the nearest neighbor and the farthest neighbor are almost equidistant in most cases in the high-dimensional data, and the dimension reduction effect is poor, so that the distance-based dimension reduction method is not applicable any more. The NMF non-negative matrix decomposition is to reduce the dimension based on the relevance between the features, the label data of the television program belongs to high-dimensional data, and the intersection of the labels in the interest field is analyzed by analyzing the relevance between the labels, so as to determine whether to merge the dimension reduction, therefore, the invention adopts the NMF non-negative matrix decomposition to compress the labels.
When the label of the special recommendation is compressed and reduced in dimension by adopting non-negative matrix decomposition, the dimension reduction principle is shown in fig. 1, a label matrix before dimension reduction in the special recommendation is a full label matrix V, the full label matrix V is an F × N order matrix, a matrix W is a scoring matrix for scoring all labels of each user, the scoring matrix W is an F × K order matrix, and H is a K × N order probability matrix, and the probability matrix H is used for reflecting label data after merging according to intersection of the labels in the interest field.
The invention relates to a television program topic recommendation method based on non-negative matrix factorization, which comprises the following steps as shown in figure 1: the method comprises the following steps of S1, obtaining label information in real time when a television program is to be recommended to a user, wherein the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information; s2, constructing a non-negative matrix decomposition model, converting the label information into a full label matrix, and performing decomposition and dimension reduction processing on the full label matrix through the non-negative matrix decomposition model to obtain a probability matrix, wherein the probability matrix is label information generated by combining according to the relevance between thematic labels; and S3, reconstructing user portrait data and thematic label data according to the label information in the probability matrix, and recommending the television program thematic according to the reconstructed user portrait data and thematic label data.
Firstly, on the basis of recommending television programs according to scores of television program labels of users in a conventional way, label information when television programs are recommended to the users is obtained, wherein the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data is used for representing scores of each label of a user, the program label data is used for representing television program information such as West shorthand, Honorameng and the like, the thematic label data is used for representing television program information of one type such as martial arts, ancient costumes, love, families and the like, label dimensions of the user portrait data, the program label data and the thematic label data are dynamically increased and updated in real time, if the dimensions of the three are respectively A, B and C, a union set of the three is taken as a label dimension, namely A U B U C, and all labels are digitally encoded according to the size of the union dimension.
And then, constructing a non-negative matrix decomposition model, converting all label information subjected to digital coding into an array full label matrix, inputting the array full label matrix into the constructed non-negative matrix decomposition model, decomposing and reducing dimensions of the full label matrix, specifically, obtaining intersections among the labels in the interest field according to analysis of the relevance among the labels, and performing fusion and dimension reduction according to the intersections to obtain corresponding scoring matrixes and probability matrixes, wherein the probability matrixes are label information generated by merging according to the relevance among the thematic labels.
The probability matrix obtained by performing decomposition and dimension reduction on the full label matrix through the nonnegative matrix decomposition model may specifically be: and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix subjected to full label matrix decomposition and dimension reduction.
And finally, recommending the TV program topics according to the user portrait data and the topic label data in the probability matrix, specifically, calculating cosine similarity between the reconstructed user portrait data and the topic label data, sequencing the cosine similarity from large to small, and selecting the TV program topics corresponding to the topic label data with the cosine similarity ranking top N for recommendation, wherein N is a positive integer greater than or equal to 1. The reconstructed thematic label data comprises new thematic label data, such as martial arts and emotional styles, wherein the martial arts comprise swordsmen and ancient costumes, and the emotional styles comprise love and family styles. And then optimizing the tags, and finally realizing reasonable distribution of the user recommendation list in the interest field through combination of the interest field.
The cosine similarity may be calculated as:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
Based on the above technical solution, the present invention further provides a television program topic recommendation system based on non-negative matrix factorization, as shown in fig. 3, including:
the system comprises an acquisition unit and a recommendation unit, wherein the acquisition unit is used for acquiring label information in real time when a television program is to be recommended to a user, and the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information;
the nonnegative matrix decomposition model is used for performing decomposition and dimension reduction processing on a full label matrix to obtain a probability matrix, the full label matrix is obtained by converting the label information, and the probability matrix is label information generated by combining the label information according to the association degree between thematic labels;
and the recommending unit is used for reconstructing the user portrait data and the thematic label data according to the label information in the probability matrix and recommending the television program thematic according to the reconstructed user portrait data and the thematic label data.
Optionally, the decomposing and dimension-reducing processing on the full label matrix to obtain the probability matrix specifically includes:
and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix after decomposition and dimension reduction processing is performed on the full label matrix.
The optional recommending unit is further configured to:
calculating cosine similarity between reconstructed user portrait data and thematic label data, sorting the cosine similarity from big to small, and selecting and recommending television program thematic corresponding to N thematic label data before the cosine similarity ranking, wherein N is a positive integer greater than or equal to 1.
Optionally, the cosine similarity calculation formula is as follows:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
It can be understood that, because the television program topic recommendation system based on non-negative matrix factorization of the present invention is a system for implementing the television program topic recommendation method based on non-negative matrix factorization, for the disclosed system, since it corresponds to the disclosed method, the description is simpler, and for the relevant points, refer to the partial description of the method. Because the television program topic recommendation method based on the non-negative matrix factorization can solve the problem that the prior television program recommendation can not reasonably cover the portrait of the user, the system for realizing the television program topic recommendation method based on the non-negative matrix factorization can also solve the problem that the prior television program recommendation can not reasonably cover the portrait of the user.
Claims (4)
1. The television program topic recommendation method based on non-negative matrix factorization is characterized by comprising the following steps of:
step 1, obtaining label information in real time when a television program is to be recommended to a user, wherein the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information;
step 2, constructing a non-negative matrix decomposition model, converting the label information into a full label matrix, and performing decomposition and dimension reduction processing on the full label matrix through the non-negative matrix decomposition model to obtain a probability matrix, wherein the probability matrix is label information generated by merging according to the relevance between thematic labels;
step 3, reconstructing user portrait data and thematic label data according to the label information in the probability matrix, and recommending the television program thematic according to the reconstructed user portrait data and thematic label data;
the recommending of the television program special topic according to the reconstructed user portrait data and the special topic tag data comprises the following steps:
calculating cosine similarity between reconstructed user portrait data and thematic label data, sorting the cosine similarity from big to small, and selecting and recommending television program thematic corresponding to N thematic label data before the cosine similarity ranking, wherein N is a positive integer greater than or equal to 1;
the cosine similarity calculation formula is as follows:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
2. The non-negative matrix factorization-based television program topic recommendation method of claim 1, wherein in step 2, the decomposing and dimension reduction processing on the full label matrix through the non-negative matrix factorization model to obtain the probability matrix specifically comprises:
and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix subjected to full label matrix decomposition and dimension reduction.
3. A television program topic recommendation system based on non-negative matrix factorization is characterized by comprising the following steps:
the system comprises an acquisition unit and a recommendation unit, wherein the acquisition unit is used for acquiring label information in real time when a television program is to be recommended to a user, and the label information comprises: the system comprises user portrait data, program label data and thematic label data, wherein the user portrait data are used for representing the scores of users on each label, the program label data are used for representing television program information, and the thematic label data are used for representing one type of television program information;
the nonnegative matrix decomposition model is used for performing decomposition and dimension reduction processing on a full label matrix to obtain a probability matrix, the full label matrix is obtained by converting the label information, and the probability matrix is label information generated by combining the label information according to the association degree between thematic labels;
the recommendation unit is used for reconstructing user portrait data and thematic label data according to the label information in the probability matrix and recommending television program thematic according to the reconstructed user portrait data and thematic label data;
the recommending of the television program special topic according to the reconstructed user portrait data and the special topic tag data comprises the following steps:
calculating cosine similarity between reconstructed user portrait data and thematic label data, sorting the cosine similarity from big to small, and selecting and recommending television program thematic corresponding to N thematic label data before the cosine similarity ranking, wherein N is a positive integer greater than or equal to 1;
the cosine similarity calculation formula is as follows:
in the formula, x represents a vector modulo length corresponding to user image data, y represents a vector modulo length corresponding to thematic tag data, and T represents transposition.
4. The non-negative matrix factorization-based television program topic recommendation system of claim 3, wherein the factorization and dimension reduction processing on the full label matrix to obtain the probability matrix is specifically as follows:
and training the non-negative matrix decomposition model by using the program label data as training data, and fitting by using the user portrait data and the special label data as fitting data to obtain a probability matrix after decomposition and dimension reduction processing is performed on the full label matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910631011.2A CN110351580B (en) | 2019-07-12 | 2019-07-12 | Television program topic recommendation method and system based on non-negative matrix factorization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910631011.2A CN110351580B (en) | 2019-07-12 | 2019-07-12 | Television program topic recommendation method and system based on non-negative matrix factorization |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110351580A CN110351580A (en) | 2019-10-18 |
| CN110351580B true CN110351580B (en) | 2021-07-13 |
Family
ID=68176122
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910631011.2A Active CN110351580B (en) | 2019-07-12 | 2019-07-12 | Television program topic recommendation method and system based on non-negative matrix factorization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110351580B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111507293A (en) * | 2020-04-22 | 2020-08-07 | 浙江工业大学 | Signal classification method based on graph convolution model |
| CN112637684B (en) * | 2020-12-25 | 2022-02-01 | 四川长虹电器股份有限公司 | Method for detecting user portrait label at smart television terminal |
| CN112732880A (en) * | 2020-12-30 | 2021-04-30 | 平安科技（深圳）有限公司 | User portrait generation method and device, computer equipment and storage medium |
| CN116017070B (en) * | 2022-12-01 | 2024-04-12 | 四川长虹电器股份有限公司 | Method for improving clicking rate of television homepage based on operation strategy |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101287082A (en) * | 2008-05-16 | 2008-10-15 | 华东师范大学 | Collaborative filtered recommendation method introducing hotness degree weight of program |
| CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
| CN103888852A (en) * | 2014-03-24 | 2014-06-25 | 清华大学 | Video recommendation method and device for social television |
| CN105282616A (en) * | 2015-11-13 | 2016-01-27 | 云南大学 | Combined collaborative filtering IPTV program recommendation method |
| CN105430505A (en) * | 2015-11-13 | 2016-03-23 | 云南大学 | IPTV program recommending method based on combined strategy |
| CN106940801A (en) * | 2016-01-04 | 2017-07-11 | 中国科学院声学研究所 | A kind of deeply for Wide Area Network learns commending system and method |
| CN108419134A (en) * | 2018-02-05 | 2018-08-17 | 华南理工大学 | The recommendation of the channels method merged with group current behavior based on individual history |
| CN108491527A (en) * | 2018-03-28 | 2018-09-04 | 百度在线网络技术（北京）有限公司 | Information recommendation method, device and electronic equipment |
| CN108920577A (en) * | 2018-06-25 | 2018-11-30 | 西北工业大学 | Television set intelligently recommended method |
| CN109597899A (en) * | 2018-09-26 | 2019-04-09 | 中国传媒大学 | The optimization method of media personalized recommendation system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8490136B2 (en) * | 2009-05-07 | 2013-07-16 | Sirius Xm Radio Inc. | Method and apparatus for providing enhanced electronic program guide with personalized selection of broadcast content using affinities data and user preferences |
| US20070199040A1 (en) * | 2006-02-23 | 2007-08-23 | Lawrence Kates | Multi-channel parallel digital video recorder |
-
2019
- 2019-07-12 CN CN201910631011.2A patent/CN110351580B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101287082A (en) * | 2008-05-16 | 2008-10-15 | 华东师范大学 | Collaborative filtered recommendation method introducing hotness degree weight of program |
| CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
| CN103888852A (en) * | 2014-03-24 | 2014-06-25 | 清华大学 | Video recommendation method and device for social television |
| CN105282616A (en) * | 2015-11-13 | 2016-01-27 | 云南大学 | Combined collaborative filtering IPTV program recommendation method |
| CN105430505A (en) * | 2015-11-13 | 2016-03-23 | 云南大学 | IPTV program recommending method based on combined strategy |
| CN106940801A (en) * | 2016-01-04 | 2017-07-11 | 中国科学院声学研究所 | A kind of deeply for Wide Area Network learns commending system and method |
| CN108419134A (en) * | 2018-02-05 | 2018-08-17 | 华南理工大学 | The recommendation of the channels method merged with group current behavior based on individual history |
| CN108491527A (en) * | 2018-03-28 | 2018-09-04 | 百度在线网络技术（北京）有限公司 | Information recommendation method, device and electronic equipment |
| CN108920577A (en) * | 2018-06-25 | 2018-11-30 | 西北工业大学 | Television set intelligently recommended method |
| CN109597899A (en) * | 2018-09-26 | 2019-04-09 | 中国传媒大学 | The optimization method of media personalized recommendation system |
Non-Patent Citations (2)
| Title |
|---|
| 《基于用户行为的社交电视推荐算法研究》;宋艳;《中国优秀硕士学位论文全文数据库》;20160615;全文 * |
| 《基于聚类与协同过滤组合算法的数字电视推荐***设计》;蓝琪;《有线电视技术》;20160815;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110351580A (en) | 2019-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110351580B (en) | Television program topic recommendation method and system based on non-negative matrix factorization | |
| Gardères et al. | Conceptbert: Concept-aware representation for visual question answering | |
| Zhao et al. | Affective image content analysis: Two decades review and new perspectives | |
| CN109992764B (en) | File generation method and device | |
| CN101346718A (en) | Method for providing user of chosen content item | |
| CN108446964B (en) | User recommendation method based on mobile traffic DPI data | |
| Yu et al. | Learning realistic facial expressions from web images | |
| CN102799637A (en) | Method for automatically generating main character abstract in television program | |
| CN111949886B (en) | Sample data generation method and related device for information recommendation | |
| CN108363748B (en) | Topic portrait system and topic portrait method based on knowledge | |
| CN105912684A (en) | Cross-media retrieval method based on visual features and semantic features | |
| CN115168745B (en) | Virtual character image recreating method and system based on image technology | |
| CN115964560B (en) | Information recommendation method and equipment based on multi-mode pre-training model | |
| CN111553742A (en) | Federal product recommendation method, device, equipment and computer storage medium | |
| CN113407696A (en) | Collection table processing method, device, equipment and storage medium | |
| CN111223014B (en) | Method and system for online generation of subdivision scene teaching courses from a large number of subdivision teaching contents | |
| CN111553743A (en) | Federal product recommendation method, device, equipment and computer storage medium | |
| CN116383363A (en) | Virtual pet chatting system | |
| KR102526263B1 (en) | Method and System for Auto Multiple Image Captioning | |
| Wu et al. | Multi-view gait recognition using NMF and 2DLDA | |
| Singh et al. | Visual perception-based criminal identification: a query-based approach | |
| CN110457590B (en) | Intelligent user portrait drawing method based on small data input | |
| CN114064969A (en) | Dynamic picture linkage display device based on emotional curve | |
| CN113761281A (en) | Virtual resource processing method, device, medium and electronic equipment | |
| Zhang et al. | Imlooking: image-based face retrieval in online dating profile search |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |