CN107277570B - A method of improving television terminal recommender system recommendation effect - Google Patents
A method of improving television terminal recommender system recommendation effect Download PDFInfo
- Publication number
- CN107277570B CN107277570B CN201710711556.5A CN201710711556A CN107277570B CN 107277570 B CN107277570 B CN 107277570B CN 201710711556 A CN201710711556 A CN 201710711556A CN 107277570 B CN107277570 B CN 107277570B
- Authority
- CN
- China
- Prior art keywords
- cold
- program
- list
- feature
- television terminal
- 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
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
-
- 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]
- H04N21/40—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
- H04N21/454—Content or additional data filtering, e.g. blocking advertisements
- H04N21/4545—Input to filtering algorithms, e.g. filtering a region of the image
-
- 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)
- Computer Graphics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a kind of methods for improving television terminal recommender system recommendation effect, recommender system is installed inside the television terminal, there is media source library, the media source library storage inside of the television terminal has all programs that can be played and its relevant information inside the television terminal.The purpose of the present invention is to provide a kind of methods for improving television terminal recommender system recommendation effect, by introducing cold list database corresponding with hot list database, cold list is constructed by media source library and user's history behavior to be recommended, effect of the recommender system on these evaluation indexes can effectively be promoted, and then improve recommendation results, recommended method disclosed by the invention can be used as a kind of independent recommended method, can also be used as a kind of supplement of other recommended methods.
Description
Technical field
The present invention relates to big data applied technical field more particularly to a kind of improvement television terminal recommender system recommendation effects
Method.
Background technique
Recommender system is obtained in e-commerce field and is answered extensively as the important tool for solving the problems, such as " information overload "
With: the project similar with items of interest to customer recommendation such as the project-based collaborative filtering recommending method of Amazon website use.
Common recommended method is divided into collaborative filtering recommending method, content-based recommendation method and mixed recommendation method in recommender system
Deng.
Actual recommendation effect, the system performance of one recommender system, can be assessed from many aspects.Common assessment refers to
Indicate: user satisfaction, explanatory, prediction accuracy, coverage rate, diversity, novelty, pleasantly surprised degree, degree of belief, real-time,
Multiple dimensions such as robustness.
Numerous recommended methods cut both ways, and the data source utilized is different, and lays particular emphasis on from data with existing not
Carry out mined information with dimension and then recommend, thus each recommended method respectively have in every common index that recommender system assess it is excellent
Bad, it is thorough to be typically difficult to.
Summary of the invention
For existing current recommender system recommendation results it is explanatory it is not strong, coverage rate is inadequate, diversity is insufficient, pleasantly surprised degree
The disadvantages of insufficient, the purpose of the present invention is to provide a kind of methods for improving television terminal recommender system recommendation effect, by drawing
Enter cold list database corresponding with hot list database, cold list is constructed by media source library and user's history behavior and is pushed away
It recommends, can effectively promote effect of the recommender system on these evaluation indexes, and then improve recommendation results, it is disclosed by the invention to push away
The method of recommending can be used as a kind of independent recommended method, can also be used as a kind of supplement of other recommended methods.
The purpose of the invention is achieved by the following technical solution:
A method of improving television terminal recommender system recommendation effect, recommendation system is installed inside the television terminal
System, the television terminal inside have media source library, and the media source library storage inside of the television terminal has and can play
All programs and its relevant information, method and step it is as follows:
A, judge whether recommender system uses for the first time;If using for the first time, then B is entered step, otherwise, enters step C;
B, cold list database corresponding with hot list database is constructed using media source library, cold list database construction method is such as
Under:
B1, program similarity matrix W is calculated according to media source libraryn*n:
In formula, wijIndicate program IiWith program IjSimilitude;
All programs of media source library are denoted as I={ I1 I2 … Ii … In, each program IiHave multiple features
Features={ feature1, feature2..., featurek..., featuref};Program can be calculated by these features
Between content similarities, and then obtain program-program similarity matrix Wn*n;
Assuming that program IiOr IjA total of f feature, wherein there is α ordinal number feature, α non-ordinal number features of f- are then saved
Mesh IiWith program IjSimilitude can be calculated by following formula:
In formula, αkThe importance of k-th of feature is indicated, if k-th of feature is ordinal number feature, program IiWith program Ij
Similitude in k-th of feature can be calculate by the following formula:
If k-th of feature is non-ordinal number feature, program IiWith program IjSimilitude in k-th of feature can pass through
Following formula calculates:
Wherein IiAnd I .featureVectorjIt .featureVector is respectively program IiWith program IjIn k-th of feature
On OneHot coding vector;
B2, according to program-program similarity matrix Wn*nIt is calculate by the following formula program IiUnexpected winner degree Coldi
ColdiIt is worth smaller, shows that this program gets over unexpected winner;To ColdiAscending order arrangement is carried out, cold list database is obtained:
ColdList=sort (Coldi)ascending;
C, cold list database corresponding with hot list database is constructed using user's viewing behavior and media source library;
C1, cold list Cold is constructed according to the method for step B using media source libraryList1;
C2, according to user's viewing behavior, count each program IiThen the number watched is pressed the number of watched time
It is arranged according to descending, obtains cold list ColdList2;
C3, fusion ColdList1And ColdList2Obtain final cold list ColdListIf ColdList2Involved in program have
P, then there is p≤n, be denoted as ColdList1={ I1 I2 … In, ColdList2={ I '1 I′2 … I′p, then it is fused
ColdListIt is obtained by following operation rule:
ColdList=(ColdList1-ColdList2)∪ColdList2
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
D, judge whether current recommender system there are other proposed algorithms, if so, being then transferred to step E, otherwise, be transferred to step F;
E, cold list database is merged with other proposed algorithms as a result, generating recommendation list RECList;
If other proposed algorithms of recommender system, other proposed algorithms include collaborative filtering or frequent item set, then other are pushed away
It recommends algorithm and generates recommendation list AlgList, then RECListIt can be obtained by following algorithm:
RECList=AlgList∪(ColdList-AlgList)
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
F: to television terminal, user recommends, by the REC of step EListOr the Cold of step CListRecommend television terminal
User.
The present invention compared with the prior art, have the following advantages that and the utility model has the advantages that
The purpose of the present invention is to provide it is a kind of improve television terminal recommender system recommendation effect method, by introduce with
The corresponding cold list database of hot list database, cold list is constructed by media source library and user's history behavior and is recommended,
Effect of the recommender system on these evaluation indexes can be effectively promoted, and then improves recommendation results, recommendation disclosed by the invention
Method can be used as a kind of independent recommended method, can also be used as a kind of supplement of other recommended methods.
Detailed description of the invention
Fig. 1 is flow diagram of the invention.
Specific embodiment
The present invention is described in further detail below with reference to embodiment:
Embodiment one
As shown in Figure 1, a kind of method for improving television terminal recommender system recommendation effect, television terminal inside installation
There is recommender system, there is media source library inside the television terminal, the media source library storage inside of the television terminal has
All programs and its relevant information that can be played, method and step are as follows:
A, judge whether recommender system uses for the first time;If using for the first time, then B is entered step, otherwise, enters step C;
B, cold list database corresponding with hot list database is constructed using media source library, cold list database construction method is such as
Under:
B1, program similarity matrix W is calculated according to media source libraryn*n:
In formula, wijIndicate program IiWith program IjSimilitude;
All programs of media source library are denoted as I={ I1 I2 … Ii … In, each program IiHave multiple features
Features={ feature1, feature2..., featurek..., featuref};Program can be calculated by these features
Between content similarities, and then obtain program-program similarity matrix Wn*n;
Assuming that program IiOr IjA total of f feature, wherein there is α ordinal number feature, α non-ordinal number features of f- are then saved
Mesh IiWith program IjSimilitude can be calculated by following formula:
In formula, αkThe importance of k-th of feature is indicated, if k-th of feature is ordinal number feature, program IiWith program Ij
Similitude in k-th of feature can be calculate by the following formula:
If k-th of feature is non-ordinal number feature, program IiWith program IjSimilitude in k-th of feature can pass through
Following formula calculates:
Wherein IiAnd I .featureVectorjIt .featureVector is respectively program IiWith program IjIn k-th of feature
On OneHot coding vector;
B2, according to program-program similarity matrix Wn*nIt is calculate by the following formula program IiUnexpected winner degree Coldi
ColdiIt is worth smaller, shows that this program gets over unexpected winner;To ColdiAscending order arrangement is carried out, cold list database is obtained:
ColdList=sort (Coldi)ascending;
C, cold list database corresponding with hot list database is constructed using user's viewing behavior and media source library;
C1, cold list Cold is constructed according to the method for step B using media source libraryList1;
C2, according to user's viewing behavior, count each program IiThen the number watched is pressed the number of watched time
It is arranged according to descending, obtains cold list ColdList2;
C3, fusion ColdList1And ColdList2Obtain final cold list ColdListIf ColdList2Involved in program have
P, then there is p≤n, be denoted as ColdList1={ I1 I2 … In, ColdList2={ I '1 I′2 … I′p, then it is fused
ColdListIt is obtained by following operation rule:
ColdList=(ColdList1-ColdList2)∪ColdList2
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
D, judge whether current recommender system there are other proposed algorithms, if so, being then transferred to step E, otherwise, be transferred to step F;
E, cold list database is merged with other proposed algorithms as a result, generating recommendation list RECList;
If other proposed algorithms of recommender system, other proposed algorithms include collaborative filtering or frequent item set, then other are pushed away
It recommends algorithm and generates recommendation list AlgList, then RECListIt can be obtained by following algorithm:
RECList=AlgList∪(ColdList-AlgList)
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
F: to television terminal, user recommends, by the REC of step EListOr the Cold of step CListRecommend television terminal
User.
Embodiment two
As shown in Figure 1, a kind of method for improving television terminal recommender system recommendation effect, television terminal inside installation
There is recommender system, there is media source library inside the television terminal, the media source library storage inside of the television terminal has
All programs and its relevant information that can be played.
Define media source library: the position for all programs and its relevant information storage that television terminal can be played claims
For media source library.
It defines user's viewing behavior: television terminal user is watched to the corelation behaviour of media resource, referred to as user's viewing row
For.
Define unexpected winner degree: the unexpected winner degree of program, i.e. unexpected winner degree, corresponding with the popular degree of program, a program is colder
Door, shows that this program is fewer and is known by public domain.
Its method and step is as follows:
Step 1: judge whether for the first time (all television terminals equipped with recommender system application do not make recommender system for use
, there is recommender system cold start-up in used recommender system, the information that can be utilized only has media source library information at this time).If
It uses for the first time, is then transferred to step 2, otherwise, be transferred to step 3.
Step 2: cold list corresponding with hot list database is constructed using content-based recommendation method using media source library
Database;
Cold list database sharing is step 1: calculate program program similarity matrix W according to media source libraryn*n:
In formula, wijIndicate program IiWith program IjSimilitude, all programs of media source library are denoted as I={ I1 I2
… Ii … In, each program IiHave multiple feature features={ feature1, feature2...,
featurek..., featuref, these features be generally divided into it is orderly and unordered, for example, the feature of motion pictures can be
{ director, performer, film duration ... }, (COUNTABLY VALUED compares) that wherein film duration is ordered into, and direct, performer is unordered.
The content similarities between program can be calculated by these features, and then obtain program-program similarity matrix Wn*n.It is now assumed that
Program IiOr IjA total of f feature, wherein there is α ordinal number feature, α non-ordinal number features of f-, then program IiWith program Ij
Similitude can be calculated by following formula:
In formula, αkIndicate the importance of k-th of feature, Ii.features (k) indicates program IiK-th of feature.
If k-th of feature is ordinal number feature, program IiWith program IjIn the case where the similitude in k-th of feature can pass through
Formula calculates:
If k-th of feature is non-ordinal number feature, program IiWith program IjSimilitude in k-th of feature can pass through
Following formula calculates:
Wherein,<,>it is inner product operation, | | | | it is 2- norm, Ii.featureVector and
IjIt .featureVector is respectively program IiWith program IjOneHot coding vector in k-th of feature.
Cold list database sharing is step 2: according to program-program similarity matrix Wn*nIt is calculate by the following formula program Ii's
Unexpected winner degree Coldi:
ColdiIt is worth smaller, shows that this program gets over unexpected winner, to ColdiCarry out the ascending order arrangement (unexpected winner of cold list database at this time
Degree is mainly shown as on the content similarities of single programs and other programs), obtain cold list: ColdList=sort
(Coldi)ascending。
Wherein, αkCan the mode of questionnaire by inquiry obtain, a kind of form of feasible questionnaire can design as follows:
For more objectively statistics, the above questionnaire should be designed to multiselect.It is now assumed that the votes of k-th of feature are nk, then αkIt can
It is calculate by the following formula:
Step 3: cold list database corresponding with " hot list " is constructed using user's viewing behavior and media source library
Cold list database sharing is step 1: construct cold list Cold using the method for step 2 using media source libraryList1;
Cold list database sharing counts each program I step 2: according to user's viewing behavioriThe number watched, then
The number of watched time is arranged according to descending, obtains cold list ColdList2, the unexpected winner degree of cold list database shows single at this time
The number that program is watched more few then this program more unexpected winner;
Cold list database sharing is step 3: fusion ColdList1And ColdList2Obtain final cold list ColdList
It is assumed that ColdList2Involved in program have p, then have p≤n, remember ColdList1={ I1 I2 … In,
ColdList2={ I '1 I′2 … I′P, then fused ColdListIt is obtained by following operation rule:
ColdList=(ColdList1-ColdList2)∪ColdList2
Wherein, symbol "-" indicates that set difference operation (and keeps Cold hereinList1The relative order of middle program is constant), symbol
Number ∪ indicates that collection union operation (and keeps (Cold hereinList1-ColdList2) and ColdList2Middle program relative order is constant, closes
And result (Cold laterList1-ColdList2) come ColdList2Before).For example, work as ColdList1={ I5 I2 I3
I1 I4 I6, ColdList2={ I2 I3 I5 I4When, (ColdList1-ColdList2)={ I1 I6, ColdList=
(ColdList1-ColdList2)∪ColdList2={ I1 I6 I2 I3 I5 I4, the meaning of above formula is if program does not appear in use
In the behavior of family, and this program with it is more dissimilar on other programme contents, then this program gets over unexpected winner.
Step 4: whether current recommender system has other proposed algorithms;If so, then going to step five, otherwise, six are gone to step.
Step 5: cold list database is merged with other proposed algorithms as a result, generating recommendation list RECList。
It is assumed that other proposed algorithms (such as collaborative filtering, frequent item set etc.) of recommender system generate recommendation list AlgList, then
RECListIt can be obtained by following algorithm:
RECList=AlgList∪(ColdList-AlgList)
In formula, each same step 3 of symbolic operation meaning, above-mentioned implication be if recommender system has other proposed algorithms,
The cold list of building supplements original proposed algorithm result with regard to the compensatory algorithm as other proposed algorithms entire to be promoted
The coverage rate of recommender system, diversity, pleasantly surprised degree and other effects index.
Step 6: to television terminal, user recommends;By RECListRecommend television terminal user.When working as step 4
Preceding recommender system does not have other proposed algorithms, directly by the Cold of step 3ListRecommend television terminal user
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (1)
1. a kind of method for improving television terminal recommender system recommendation effect, it is characterised in that: installation inside the television terminal
There is recommender system, there is media source library inside the television terminal, the media source library storage inside of the television terminal has
All programs and its relevant information that can be played, method and step are as follows:
A, judge whether recommender system uses for the first time;If using for the first time, then B is entered step, otherwise, enters step C;
B, cold list database corresponding with hot list database is constructed using media source library, cold list database construction method is as follows:
B1, program similarity matrix W is calculated according to media source libraryn*n:
In formula, wijIndicate program IiWith program IjSimilitude;
All programs of media source library are denoted as I={ I1 I2 … Ii … In, each program IiHave multiple features
Features={ feature1,feature2,…,featurek,…,featuref};Program can be calculated by these features
Between content similarities, and then obtain program-program similarity matrix Wn*n;
Assuming that program IiOr IjA total of f feature, wherein there is α ordinal number feature, α non-ordinal number features of f-, then program Ii
With program IjSimilitude can be calculated by following formula:
In formula, αkThe importance of k-th of feature is indicated, if k-th of feature is ordinal number feature, program IiWith program IjIn kth
Similitude in a feature can be calculate by the following formula:
If k-th of feature is non-ordinal number feature, program IiWith program IjSimilitude in k-th of feature can pass through following formula
It calculates:
Wherein IiAnd I .featureVectorjIt .featureVector is respectively program IiWith program IjIn k-th of feature
OneHot coding vector;
B2, according to program-program similarity matrix Wn*nIt is calculate by the following formula program IiUnexpected winner degree Coldi
ColdiIt is worth smaller, shows that this program gets over unexpected winner;To ColdiAscending order arrangement is carried out, cold list database: Cold is obtainedList=
sort(Coldi)ascending;
C, cold list database corresponding with hot list database is constructed using user's viewing behavior and media source library;
C1, cold list Cold is constructed according to the method for step B using media source libraryList1;
C2, according to user's viewing behavior, count each program IiThe number watched, then by the number of watched time according to drop
Sequence arrangement, obtains cold list ColdList2;
C3, fusion ColdList1And ColdList2Obtain final cold list ColdListIf ColdList2Involved in program have p,
Then there is p≤n, is denoted as
ColdList1={ I1 I2 … In, ColdList2={ I '1 I′2 … I′p, then fused ColdListBy following
Operation rule obtains:
ColdList=(ColdList1-ColdList2)∪ColdList2
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
D, judge whether current recommender system there are other proposed algorithms, if so, being then transferred to step E, otherwise, be transferred to step F;
E, cold list database is merged with other proposed algorithms as a result, generating recommendation list RECList;
If other proposed algorithms of recommender system, other proposed algorithms include collaborative filtering or frequent item set, then other recommend to calculate
Method generates recommendation list AlgList, then RECListIt can be obtained by following algorithm:
RECList=AlgList∪(ColdList-AlgList)
Wherein, symbol "-" indicates that set difference operation, symbol ∪ indicate collection union operation;
F: to television terminal, user recommends, by the REC of step EListOr the Cold of step CListRecommend television terminal use
Family.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710711556.5A CN107277570B (en) | 2017-08-18 | 2017-08-18 | A method of improving television terminal recommender system recommendation effect |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710711556.5A CN107277570B (en) | 2017-08-18 | 2017-08-18 | A method of improving television terminal recommender system recommendation effect |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107277570A CN107277570A (en) | 2017-10-20 |
CN107277570B true CN107277570B (en) | 2019-11-05 |
Family
ID=60080412
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710711556.5A Active CN107277570B (en) | 2017-08-18 | 2017-08-18 | A method of improving television terminal recommender system recommendation effect |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107277570B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109618231B (en) * | 2018-11-15 | 2021-07-27 | 深圳市酷开网络科技股份有限公司 | Movie content recommendation method and system based on cold start |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1538838A1 (en) * | 2003-12-02 | 2005-06-08 | Sony Corporation | Information processor, information processing method and computer program |
CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
CN102780920A (en) * | 2011-07-05 | 2012-11-14 | 上海奂讯通信安装工程有限公司 | Television program recommending method and system |
CN103546778A (en) * | 2013-07-17 | 2014-01-29 | Tcl集团股份有限公司 | Television program recommendation method and system, and implementation method of system |
CN105404700A (en) * | 2015-12-30 | 2016-03-16 | 山东大学 | Collaborative filtering-based video program recommendation system and recommendation method |
-
2017
- 2017-08-18 CN CN201710711556.5A patent/CN107277570B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1538838A1 (en) * | 2003-12-02 | 2005-06-08 | Sony Corporation | Information processor, information processing method and computer program |
CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
CN102780920A (en) * | 2011-07-05 | 2012-11-14 | 上海奂讯通信安装工程有限公司 | Television program recommending method and system |
CN103546778A (en) * | 2013-07-17 | 2014-01-29 | Tcl集团股份有限公司 | Television program recommendation method and system, and implementation method of system |
CN105404700A (en) * | 2015-12-30 | 2016-03-16 | 山东大学 | Collaborative filtering-based video program recommendation system and recommendation method |
Also Published As
Publication number | Publication date |
---|---|
CN107277570A (en) | 2017-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103686382B (en) | Program recommendation method | |
CN104317835B (en) | The new user of video terminal recommends method | |
Nóbrega et al. | Towards explaining recommendations through local surrogate models | |
Kwon et al. | Personalized smart TV program recommender based on collaborative filtering and a novel similarity method | |
JP2004524764A (en) | Method and apparatus for generating recommended scores | |
CN104053023B (en) | A kind of method and device of determining video similarity | |
CN105430504A (en) | Family member mix identification method and system based on television watching log mining | |
US20140259040A1 (en) | Methods and apparatuses for deducing a viewing household member profile | |
CN104394471A (en) | Method for intelligently recommending favorite program to user | |
Ullah et al. | Hybrid recommender system with temporal information | |
Dragone et al. | Deriving user-and content-specific rewards for contextual bandits | |
Aharon et al. | Watch-it-next: a contextual TV recommendation system | |
CN107277570B (en) | A method of improving television terminal recommender system recommendation effect | |
CN104008193A (en) | Information recommending method based on typical user group finding technique | |
Rathod et al. | A survey of personalized recommendation system with user interest in social network | |
US20160094887A1 (en) | Tv program-based shopping guide system and tv program-based shopping guide method | |
Wang et al. | Towards effective recommendation of social data across social networking sites | |
Nikolakopoulos et al. | A novel hierarchical approach to ranking-based collaborative filtering | |
Shah et al. | A hybrid based recommendation system based on clustering and association | |
Hu et al. | Collaborative filtering recommendation algorithm based on user explicit preference | |
CN111506816B (en) | Recommendation method, recommendation device, recommendation equipment and storage medium | |
Padmanabhan et al. | A rule based approach to group recommender systems | |
Rahman | Contextual recommender systems using a multidimensional approach | |
CN105472460A (en) | Intelligent sorting method and system of television applications | |
Poompuang et al. | User and item pattern matching in multi-criteria recommender systems |
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 |