CN103514239A - Recommendation method and system integrating user behaviors and object content - Google Patents
Recommendation method and system integrating user behaviors and object content Download PDFInfo
- Publication number
- CN103514239A CN103514239A CN201210486275.1A CN201210486275A CN103514239A CN 103514239 A CN103514239 A CN 103514239A CN 201210486275 A CN201210486275 A CN 201210486275A CN 103514239 A CN103514239 A CN 103514239A
- Authority
- CN
- China
- Prior art keywords
- user
- article
- content
- data
- recommendation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
- G06F16/337—Profile generation, learning or modification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a recommendation method and system integrating user behaviors and object content. The recommendation method integrating the user behaviors and the object content comprises the steps that through correlation between behavior data of each user and content data of all the objects, a list of user interest of each user to all the objects is obtained; the similarity of the list of user interest and the content data of all the objects is calculated, the recommendation object weight of each user to the content data of each object is obtained, and the objects recommended to each user are obtained through ranking of the recommendation object weights. According to the recommendation method and system integrating the user behaviors and the object content, the main problems that how an existing recommendation conducts cold start and how the existing recommendation smoothly transmits to the normal operation state from cold start are solved, the accuracy, the coverage rate, the novelty and the like of the system are greatly improved, the accuracy and the personalization of the recommendation system can be well improved, and more users are attracted and use the recommendation system.
Description
Technical field
The present invention relates to intelligent recommendation system, in particular recommend method and the system of a kind of integrated user behavior and article content.
Background technology
Along with the development of infotech and internet, how from a large amount of information, to find the interested information of user, how to allow oneself information be subject to users' welcome, for user and provider, be all a very difficult thing.The task of commending system is exactly contact user and information, helps on the one hand user to find own valuable information, and allows on the other hand information can be presented in face of its interesting user, thereby realize information consumer and informant's doulbe-sides' victory.Commending system is mainly the behavior by analysis user, to its modeling, thus the interest of predictive user recommending to user.
According to the different modes of deal with data, the method that commending system is used mainly can be divided into collaborative filtering and information filtering, if commending system only utilizes user's behavioral data, according to user's historical interest, to user, recommend, so this method is called as collaborative filtering (collaborative filtering).If commending system has utilized the content-data (content filtering) of article, calculate user's interest and the similarity between article description, to user, recommend, be called information filtering.Collaborative filtering is the foremost algorithm in commending system field, this algorithm mainly the historical behavior by research user to user's interest modeling and make recommendation to user.It mainly comprises the collaborative filtering based on user, the collaborative filtering based on article, and the model based on theme.Information filtering is to user, to recommend and the article of liking before them similar other article in terms of content on the basis based on article content.Content mainly comprises some attributes of article.This method can be well understood to user's point of interest more, thereby better explains the reason of recommending.
Collaborative filtering mainly depends on historical behavior data, and the therefore good cold start-up problem of resolution system that is to say that these algorithms cannot recommend to new user, because there is no relevant behavioral data.Also new article cannot be recommended may be to its interested user simultaneously.Information filtering algorithm is mainly to utilize the similarity of content between article to recommend to user, and it only depends on the data of article content itself, so can solve cold start-up problem.But because information filtering algorithm has been ignored user behavior, thereby also ignored the rule comprising in the popularity of article and user behavior, so its ratio of precision collaborative filtering is low.The advantage that how to merge these two kinds of algorithms designs a kind of new proposed algorithm cold start-up facing for commending system and the precision that improves commending system and is of great significance.
Therefore, prior art has yet to be improved and developed.
Summary of the invention
The technical problem to be solved in the present invention is, above-mentioned defect for prior art, recommend method and the system of a kind of integrated user behavior and article content are provided, solve the cold start-up of existing commending system and how from cold start-up smoothly excessively to the problem of normal operating condition.
The technical scheme that technical solution problem of the present invention adopts is as follows:
A recommend method for integrated user behavior and article content, wherein, comprises the following steps:
A, by associated each user's behavioral data and the content-data of all article, obtain the user interest list of each user to all article contents;
B, calculate the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
Described integrated user behavior and the recommend method of article content, wherein, described steps A specifically comprises:
A1, respectively the content-data of user's behavioral data and article is adopted to user's behavior vector and the content vector representation of article, and the content vector of the behavior vector sum article of the method associated user by matrix decomposition, obtain user's interest vector.
Described integrated user behavior and the recommend method of article content, wherein, described step B specifically also comprises:
By the content-data of all article, upgrade the content-data of newly-increased article, and by calculating the similarity of the content-data of described newly-increased article and all users' behavioral data, draw the recommendation user weight of newly-increased article to user's behavioral data, and by draw the recommendation user of newly-increased article to recommending article weight to sort.
Described integrated user behavior and the recommend method of article content, wherein, calculate user's interest vector by following formula:
Wherein,
the behavior vector u that represents unique user,
the content vector c that represents Individual Items,
the content matrix C that represents all article,
the interest vector f that represents unique user, n and m are natural number,
;
When the content matrix of all article is fixedly time, by above-mentioned formula (1), derive the interest vector f of unique user.
Described integrated user behavior and the recommend method of article content, wherein, when Add User and described in Add User while producing behavioral data, the behavior vector Adding User by the behavioral data Adding User, and draw by above-mentioned formula (1) interest vector Adding User, and derive by following formula (2) the recommendation article weight vectors Adding User:
Wherein,
represent to recommend article weight vectors, by above-mentioned recommendation article weight vectors, obtain and recommend article to recommend user.
Described integrated user behavior and the recommend method of article content, wherein, when having newly-increased article, upgrade the content vector of newly-increased article, and by formula (3), derive the recommendation user weight vectors of newly-increased article according to the content matrix C of described all article:
formula (3)
Wherein,
represent described user behavior matrix U,
represent to recommend user's weight vectors, by above-mentioned recommendation user weight vectors, obtain the recommendation user of newly-increased article.
Described integrated user behavior and the recommend method of article content, wherein, described steps A also comprises: the interest vector that adopts the data structure calculating unique user of sparse storage.
A commending system for integrated user behavior and article content, wherein, described system comprises:
Relating module, for by associated each user's behavioral data and the content-data of all article, obtains the user interest list of each user to all article contents;
Recommend article module, for calculating the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
Described integrated user behavior and the commending system of article content, wherein, described system comprises:
Recommend line module, for the content-data by all article, upgrade the content-data of newly-increased article, and by calculating the similarity of the content-data of described newly-increased article and all users' behavioral data, draw the recommendation user weight of newly-increased article to user's behavioral data, and by draw the recommendation user of newly-increased article to recommending article weight to sort.
Recommend method and the system of integrated user behavior provided by the present invention and article content, effectively solved the subject matter that current commending system faces, cold start-up such as commending system, and how from cold start-up smoothly excessively to normal operating condition etc., rate of accurateness, coverage rate, novel degree etc. have greatly been improved, can better improve precision and the personalization of commending system, attract more user to use.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the recommend method of integrated user behavior provided by the invention and article content.
Fig. 2 is the structural representation of the commending system of integrated user behavior provided by the invention and article content.
Embodiment
For making object of the present invention, technical scheme and advantage clearer, clear and definite, referring to accompanying drawing, developing simultaneously, the present invention is described in more detail for embodiment.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.
Referring to Fig. 1, Fig. 1 is the process flow diagram of the recommend method of integrated user behavior provided by the invention and article content, comprises the following steps:
Step S100, by associated each user's behavioral data and the content-data of all article, obtain the user interest list of each user to all article contents;
Step S200, calculate the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
Below in conjunction with specific embodiment, above-mentioned steps is described in detail.
The present invention carrys out the interest modeling to user by the content-data of integrated article and user's behavioral data, thereby effectively in conjunction with the advantage of information filtering and collaborative filtering, user is recommended.Particularly, the present invention is mainly divided into off-line and online two parts, off-line part is mainly in order to obtain user's interest list, online part is the user interest list by partly obtaining at off-line, provides its corresponding recommendation article and when newly-increased article, provide its corresponding user of recommendation when Adding User.
In off-line part, in order to obtain user's interest list, first user's interest is carried out to modeling, define a topic model, in order to obtain user's interest list.By the content-data of associated article and user's behavioral data, topic model utilizes the data mining algorithm of matrix decomposition to obtain the interest list of user to article content.At off-line, partly obtain user as follows to the main flow process of the interest list of article content: first respectively the content-data of user's behavioral data and article is adopted to user's behavior vector and the content vector representation of article, the behavior vector u of unique user is expressed as
,
All users' behavioural matrix U=
, the content vector c of Individual Items is expressed as
, the content matrix of all article
, the interest vector f of unique user uses
represent, topic model is like this:
The content-data of existing subscriber's behavioral data and article is obtained to user's interest list by topic model; And if have new user behavior or an input of new article content, upgrade the vector space model of all user behavior datas or all article content-data, to obtain more accurately the weight of its feature in overall vector space model, then by topic model, go to obtain the user interest list after renewal like this.
In online part, set up the recommended models of commending system, recommended models is mainly weighed the fancy grade of user to each article by user's interest list, and provides online recommendation results.Here, recommended models is divided into two types, and one is the recommendation article model for user, and one is the recommendation user model for article.When recommended models is defined, according to the difference of recommended models type, carry out different processing.When user is recommended to article, if existing user, can directly inquire this user's interest list, then by recommending article model to recommend online article; If Add User and produced corresponding behavioral data, in order to obtain exactly the weight of this behavioral data Adding User in the overall situation, first the behavioral data by all users removes to upgrade the behavioral data that this Adds User, and then by topic model, is gone to obtain its user interest list and is provided recommendation article.When newly-increased article are recommended user, if there is the content-data of newly-increased article, first by all the elements data, remove to upgrade the content-data of these newly-increased article, then by all users' interest list and recommendation user model, go to obtain recommendation user.
Particularly, recommend article model as follows:
Recommendation user model is as follows:
Concrete recommendation process is: when Add User and described in Add User while producing behavioral data, the behavior vector Adding User by the behavioral data Adding User, and by drawing the interest vector Adding User, and the recommendation article weight vectors that goes out to Add User by following recommendation article model inference, is obtained and is recommended article to recommend user by above-mentioned recommendation article weight vectors.When having newly-increased article, according to the content matrix C of described all article, upgrade the content vector of newly-increased article, and by recommending user model to derive the recommendation user weight vectors of newly-increased article, then by above-mentioned recommendation user weight vectors, obtain the recommendation user of newly-increased article.
By above-mentioned two recommended models, can draw respectively the weight vectors of recommending article and the weight vectors of recommending user, according to recommending the weight vectors of article and recommending user's weight vectors can provide very accurately recommendation article and recommend user, the recommendation article here and recommendation user are not limited to recommend one, can recommend a plurality of.By the hobby of interest list analysis user, draw good rationale for the recommendation.
The present invention adopts matrix decomposition method to obtain user's interest list, and concrete computation process is as follows: the content-data of article and user's behavioral data represent with vector space model.The content weight vectors of each article can be expressed as
, each user's behavior vector can be expressed as
, can be by TF-IDF(Term Frequency – Inverse Document Frequency, the anti-document frequency of Ci Pin –) and method calculates, the error to avoid calculating by these vectorial regularizations simultaneously;
TF(word frequency) number of times that calculated characteristics occurs in the vector of current object, DF(document frequency) calculating feature in whole vector space appears in how many different objects.When having new user behavior data or new article content-data to add, will upgrade according to this computation model.
Definition user's interests matrix is F, in order to obtain F, and objective definition function J:
By using Lagrangian function to solve this condition objective function J.Because
matrix is fixing, therefore a demand solution
.Suppose that its corresponding Lagrange multiplier is
, LagrangianL is defined as:
Therefore,
From above-mentioned formula, can find out, objective function is monotone decreasing convergence.When initialization,
can be some random numbers, but in order to obtain unique user interest list, our initialization
for being 1 matrix entirely, that is to say that we think that user is consistent to all interest before obtaining real user interest vector.Along with the constantly iteration of algorithm,
little by little convergence.The end condition of algorithm can arrange certain iterations, or the error precision of working as objective function is within certain scope.
The number of simultaneously considering user and article is very huge, and user behavior data and article content-data matrix are very sparse, in order to improve the speed of calculating, reduce the consumption of internal memory simultaneously, adopt sparse storage mode, data structure by sparse storage is calculated the interest vector of unique user, is about to the value of non-zero and its and quotes two arrays and store, and can effectively carry out the computing of sparse matrix like this, solve large matrix computing, obtain rapidly recommendation list.
At the above-mentioned unique user of giving
while recommending article, need to calculate user's interest vector
content matrix with all article
between recommendation article weight vectors
, due to
with
all immobilizations, so their product cosine similarity namely.By recommendation article weight vectors is sorted to obtain maximally related article.In like manner, in the time increasing article newly and recommend user, calculate the content vector of newly-increased article
interests matrix with all users
between recommendation user weight vectors
.Then it being sorted to obtain the user of high several weights recommends.
Further, the behavioral data of the user in the present invention can have multiple expression-form, such as dominant feedback or stealthy feedback, shows that feedback can be that user is to the scoring of article and stealthy feedback can browsing record and represent from user.And the content-data of article can be used the label of article, textual description of article etc.Which kind of representation no matter, the model that the present invention proposes takes full advantage of various data representation form and recommends, and the whole process of disposal system from cold start-up to normal operation effectively, thereby has practicality widely.
Based on above-mentioned recommend method, the present invention also provides the commending system of a kind of integrated user behavior and article content, and as shown in Figure 2, described system comprises:
Relating module 10, for by associated each user's behavioral data and the content-data of all article, obtains the user interest list of each user to all article contents;
Recommend article module 20, for calculating the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
Recommend line module 30, for the content-data by all article, upgrade the content-data of newly-increased article, and by calculating the similarity of the content-data of described newly-increased article and all users' behavioral data, draw the recommendation user weight of newly-increased article to user's behavioral data, and by draw the recommendation user of newly-increased article to recommending article weight to sort.
Preferably, commending system provided by the invention can be widely used in various intelligent recommendation systems, such as film commending system, music recommend system, reading commending system etc.
In sum, recommend method and the system of integrated user behavior provided by the invention and article content, by data mining model effectively in conjunction with the advantage of this behavior recommendation and commending contents, effectively solve the subject matter that current commending system faces, such as the cold start-up of commending system, and how from cold start-up smoothly excessively to normal operating condition etc.The recommend method of integrated user behavior of the present invention and article content and system can overcome not to be had content or there is no these two kinds of special circumstances of behavior, and rate of accurateness, coverage rate and novel degree etc. have greatly been improved, can better improve precision and the personalization of commending system, attract more user to use.
Should be understood that, application of the present invention is not limited to above-mentioned giving an example, and for those of ordinary skills, can be improved according to the above description or convert, and all these improvement and conversion all should belong to the protection domain of claims of the present invention.
Claims (9)
1. a recommend method for integrated user behavior and article content, is characterized in that, comprises the following steps:
A, by associated each user's behavioral data and the content-data of all article, obtain the user interest list of each user to all article contents;
B, calculate the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
2. the recommend method of integrated user behavior according to claim 1 and article content, is characterized in that, described steps A specifically comprises:
A1, respectively the content-data of user's behavioral data and article is adopted to user's behavior vector and the content vector representation of article, and the content vector of the behavior vector sum article of the method associated user by matrix decomposition, obtain user's interest vector.
3. the recommend method of integrated user behavior according to claim 1 and article content, is characterized in that, described step B specifically also comprises:
By the content-data of all article, upgrade the content-data of newly-increased article, and by calculating the similarity of the content-data of described newly-increased article and all users' behavioral data, draw the recommendation user weight of newly-increased article to user's behavioral data, and by draw the recommendation user of newly-increased article to recommending article weight to sort.
4. the recommend method of integrated user behavior according to claim 2 and article content, is characterized in that, calculates user's interest vector by following formula:
Wherein,
the behavior vector u that represents unique user,
the content vector c that represents Individual Items,
the content matrix C that represents all article,
the interest vector f that represents unique user, n and m are natural number,
;
When the content matrix of all article is fixedly time, by above-mentioned formula (1), derive the interest vector f of unique user.
5. the recommend method of integrated user behavior according to claim 4 and article content, it is characterized in that, when Add User and described in Add User while producing behavioral data, the behavior vector Adding User by the behavioral data Adding User, and draw by above-mentioned formula (1) interest vector Adding User, and derive by following formula (2) the recommendation article weight vectors Adding User:
6. the recommend method of integrated user behavior according to claim 4 and article content, it is characterized in that, when having newly-increased article, according to the content matrix C of described all article, upgrade the content vector of newly-increased article, and by formula (3), derive the recommendation user weight vectors of newly-increased article:
formula (3)
7. the recommend method of integrated user behavior according to claim 2 and article content, is characterized in that, described steps A also comprises: the interest vector that adopts the data structure calculating unique user of sparse storage.
8. a commending system for integrated user behavior and article content, is characterized in that, described system comprises:
Relating module, for by associated each user's behavioral data and the content-data of all article, obtains the user interest list of each user to all article contents;
Recommend article module, for calculating the similarity of the content-data of described user interest list and described all article, draw the recommendation article weight of each user to the content-data of article, and by drawing each user's recommendation article to recommending article weight to sort.
9. the commending system of integrated user behavior according to claim 8 and article content, is characterized in that, described system comprises:
Recommend line module, for the content-data by all article, upgrade the content-data of newly-increased article, and by calculating the similarity of the content-data of described newly-increased article and all users' behavioral data, draw the recommendation user weight of newly-increased article to user's behavioral data, and by draw the recommendation user of newly-increased article to recommending article weight to sort.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210486275.1A CN103514239B (en) | 2012-11-26 | 2012-11-26 | A kind of integrated user behavior and the recommendation method and system of item contents |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210486275.1A CN103514239B (en) | 2012-11-26 | 2012-11-26 | A kind of integrated user behavior and the recommendation method and system of item contents |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103514239A true CN103514239A (en) | 2014-01-15 |
CN103514239B CN103514239B (en) | 2016-12-21 |
Family
ID=49896970
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210486275.1A Active CN103514239B (en) | 2012-11-26 | 2012-11-26 | A kind of integrated user behavior and the recommendation method and system of item contents |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103514239B (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104063589A (en) * | 2014-06-16 | 2014-09-24 | 百度移信网络技术(北京)有限公司 | Recommendation method and system |
CN104766219A (en) * | 2015-03-19 | 2015-07-08 | 中国船舶重工集团公司第七0九研究所 | User recommendation list generation method and system based on taking list as unit |
CN105574025A (en) * | 2014-10-15 | 2016-05-11 | 阿里巴巴集团控股有限公司 | Methods and devices for sorting score calculation and model building, and commodity recommendation system |
CN105787055A (en) * | 2016-02-26 | 2016-07-20 | 合网络技术(北京)有限公司 | Information recommendation method and device |
CN106326277A (en) * | 2015-06-30 | 2017-01-11 | 上海证大喜马拉雅网络科技有限公司 | User behavior-based personalized audio recommendation method and system |
CN106599047A (en) * | 2016-11-10 | 2017-04-26 | 咪咕动漫有限公司 | Information pushing method and device |
CN106651542A (en) * | 2016-12-31 | 2017-05-10 | 珠海市魅族科技有限公司 | Goods recommendation method and apparatus |
CN106874374A (en) * | 2016-12-31 | 2017-06-20 | 杭州益读网络科技有限公司 | A kind of recommendation method for pushing based on user's history behavior interaction analysis |
CN106959966A (en) * | 2016-01-12 | 2017-07-18 | 腾讯科技(深圳)有限公司 | A kind of information recommendation method and system |
CN107239993A (en) * | 2017-05-24 | 2017-10-10 | 海南大学 | A kind of matrix decomposition recommendation method and system based on expansion label |
CN107665208A (en) * | 2016-07-28 | 2018-02-06 | 北京国双科技有限公司 | User preference measure and device |
CN108038217A (en) * | 2017-12-22 | 2018-05-15 | 北京小度信息科技有限公司 | Information recommendation method and device |
CN109389497A (en) * | 2018-09-05 | 2019-02-26 | 平安科技(深圳)有限公司 | Intelligent recommendation method, apparatus, computer equipment and the storage medium of product sale |
CN109753275A (en) * | 2018-12-10 | 2019-05-14 | 东软集团股份有限公司 | Recommended method, device, storage medium and the electronic equipment of Application Programming Interface |
CN109829116A (en) * | 2019-02-14 | 2019-05-31 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, server and computer readable storage medium |
CN110008408A (en) * | 2019-04-12 | 2019-07-12 | 山东大学 | A kind of session recommended method, system, equipment and medium |
CN110223107A (en) * | 2019-05-23 | 2019-09-10 | 中国银行股份有限公司 | Method, apparatus and equipment are determined based on the reference advertisement of analogical object |
CN112312216A (en) * | 2020-10-16 | 2021-02-02 | 山东海看新媒体研究院有限公司 | Traceable television recommendation method and system based on modular factorial theory |
CN112559856A (en) * | 2020-12-04 | 2021-03-26 | 中国联合网络通信集团有限公司 | Application sorting method and device |
CN113744021A (en) * | 2021-02-08 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Recommendation method, recommendation device, computer storage medium and recommendation system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101673286A (en) * | 2008-09-08 | 2010-03-17 | 索尼株式会社 | Apparatus, method and computer program for content recommendation and recording medium |
CN101923545A (en) * | 2009-06-15 | 2010-12-22 | 北京百分通联传媒技术有限公司 | Method for recommending personalized information |
CN102426686A (en) * | 2011-09-29 | 2012-04-25 | 南京大学 | Internet information product recommending method based on matrix decomposition |
WO2012093046A2 (en) * | 2011-01-05 | 2012-07-12 | Thomson Licensing | Hybrid content recommendation system using matrices breakdowns |
-
2012
- 2012-11-26 CN CN201210486275.1A patent/CN103514239B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101673286A (en) * | 2008-09-08 | 2010-03-17 | 索尼株式会社 | Apparatus, method and computer program for content recommendation and recording medium |
CN101923545A (en) * | 2009-06-15 | 2010-12-22 | 北京百分通联传媒技术有限公司 | Method for recommending personalized information |
WO2012093046A2 (en) * | 2011-01-05 | 2012-07-12 | Thomson Licensing | Hybrid content recommendation system using matrices breakdowns |
CN102426686A (en) * | 2011-09-29 | 2012-04-25 | 南京大学 | Internet information product recommending method based on matrix decomposition |
Non-Patent Citations (2)
Title |
---|
PATRICK BAUDISCH: "Joining Collaborative and Content-based Filtering", 《THE ACM CHI WORKSHOPON INTERACTING WITH RECOMMENDER SYSTEMS》, 31 May 1999 (1999-05-31), pages 1 - 4 * |
郭艳红 等: "协同过滤系统项目冷启动的混合推荐算法", 《计算机工程》, vol. 34, no. 23, 31 December 2008 (2008-12-31), pages 11 - 13 * |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104063589A (en) * | 2014-06-16 | 2014-09-24 | 百度移信网络技术(北京)有限公司 | Recommendation method and system |
CN104063589B (en) * | 2014-06-16 | 2018-01-16 | 百度移信网络技术(北京)有限公司 | A kind of recommendation method and system |
CN105574025A (en) * | 2014-10-15 | 2016-05-11 | 阿里巴巴集团控股有限公司 | Methods and devices for sorting score calculation and model building, and commodity recommendation system |
CN105574025B (en) * | 2014-10-15 | 2018-10-16 | 阿里巴巴集团控股有限公司 | For calculating sequence point and establishing the method, apparatus and commercial product recommending system of model |
CN104766219B (en) * | 2015-03-19 | 2018-05-22 | 中国船舶重工集团公司第七0九研究所 | Based on the user's recommendation list generation method and system in units of list |
CN104766219A (en) * | 2015-03-19 | 2015-07-08 | 中国船舶重工集团公司第七0九研究所 | User recommendation list generation method and system based on taking list as unit |
CN106326277A (en) * | 2015-06-30 | 2017-01-11 | 上海证大喜马拉雅网络科技有限公司 | User behavior-based personalized audio recommendation method and system |
CN106326277B (en) * | 2015-06-30 | 2019-07-16 | 上海证大喜马拉雅网络科技有限公司 | A kind of audio personalization recommended method and system based on user behavior |
CN106959966A (en) * | 2016-01-12 | 2017-07-18 | 腾讯科技(深圳)有限公司 | A kind of information recommendation method and system |
CN105787055A (en) * | 2016-02-26 | 2016-07-20 | 合网络技术(北京)有限公司 | Information recommendation method and device |
CN107665208B (en) * | 2016-07-28 | 2019-12-13 | 北京国双科技有限公司 | User preference measurement method and device |
CN107665208A (en) * | 2016-07-28 | 2018-02-06 | 北京国双科技有限公司 | User preference measure and device |
CN106599047A (en) * | 2016-11-10 | 2017-04-26 | 咪咕动漫有限公司 | Information pushing method and device |
CN106874374A (en) * | 2016-12-31 | 2017-06-20 | 杭州益读网络科技有限公司 | A kind of recommendation method for pushing based on user's history behavior interaction analysis |
CN106651542A (en) * | 2016-12-31 | 2017-05-10 | 珠海市魅族科技有限公司 | Goods recommendation method and apparatus |
CN107239993A (en) * | 2017-05-24 | 2017-10-10 | 海南大学 | A kind of matrix decomposition recommendation method and system based on expansion label |
CN107239993B (en) * | 2017-05-24 | 2020-11-24 | 海南大学 | Matrix decomposition recommendation method and system based on expansion label |
CN108038217A (en) * | 2017-12-22 | 2018-05-15 | 北京小度信息科技有限公司 | Information recommendation method and device |
CN108038217B (en) * | 2017-12-22 | 2021-05-11 | 北京星选科技有限公司 | Information recommendation method and device |
CN109389497A (en) * | 2018-09-05 | 2019-02-26 | 平安科技(深圳)有限公司 | Intelligent recommendation method, apparatus, computer equipment and the storage medium of product sale |
WO2020048062A1 (en) * | 2018-09-05 | 2020-03-12 | 平安科技(深圳)有限公司 | Intelligent recommendation method and apparatus for product sales, computer device and storage medium |
CN109753275A (en) * | 2018-12-10 | 2019-05-14 | 东软集团股份有限公司 | Recommended method, device, storage medium and the electronic equipment of Application Programming Interface |
CN109753275B (en) * | 2018-12-10 | 2022-04-12 | 东软集团股份有限公司 | Recommendation method and device for application programming interface, storage medium and electronic equipment |
CN109829116A (en) * | 2019-02-14 | 2019-05-31 | 北京达佳互联信息技术有限公司 | A kind of content recommendation method, device, server and computer readable storage medium |
CN109829116B (en) * | 2019-02-14 | 2021-07-30 | 北京达佳互联信息技术有限公司 | Content recommendation method and device, server and computer readable storage medium |
CN110008408A (en) * | 2019-04-12 | 2019-07-12 | 山东大学 | A kind of session recommended method, system, equipment and medium |
CN110008408B (en) * | 2019-04-12 | 2021-04-06 | 山东大学 | Session recommendation method, system, device and medium |
CN110223107A (en) * | 2019-05-23 | 2019-09-10 | 中国银行股份有限公司 | Method, apparatus and equipment are determined based on the reference advertisement of analogical object |
CN112312216A (en) * | 2020-10-16 | 2021-02-02 | 山东海看新媒体研究院有限公司 | Traceable television recommendation method and system based on modular factorial theory |
CN112312216B (en) * | 2020-10-16 | 2022-08-16 | 山东海看新媒体研究院有限公司 | Traceable television recommendation method and system based on modular factorial theory |
CN112559856A (en) * | 2020-12-04 | 2021-03-26 | 中国联合网络通信集团有限公司 | Application sorting method and device |
CN112559856B (en) * | 2020-12-04 | 2023-06-02 | 中国联合网络通信集团有限公司 | Application ordering method and device |
CN113744021A (en) * | 2021-02-08 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Recommendation method, recommendation device, computer storage medium and recommendation system |
Also Published As
Publication number | Publication date |
---|---|
CN103514239B (en) | 2016-12-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103514239A (en) | Recommendation method and system integrating user behaviors and object content | |
Li et al. | Learning from history and present: Next-item recommendation via discriminatively exploiting user behaviors | |
Xu et al. | A social-aware service recommendation approach for mashup creation | |
CN102208086B (en) | Field-oriented personalized intelligent recommendation system and implementation method | |
Abdollahi et al. | Explainable restricted boltzmann machines for collaborative filtering | |
CN106600302A (en) | Hadoop-based commodity recommendation system | |
CN103399858A (en) | Socialization collaborative filtering recommendation method based on trust | |
CN103678672A (en) | Method for recommending information | |
CN104063481A (en) | Film individuation recommendation method based on user real-time interest vectors | |
Ahmed et al. | Latent factor models with additive and hierarchically-smoothed user preferences | |
CN105787767A (en) | Method and system for obtaining advertisement click-through rate pre-estimation model | |
Liu et al. | Towards a dynamic top-n recommendation framework | |
Wang et al. | CAMO: A collaborative ranking method for content based recommendation | |
Yang et al. | A group preference-based item similarity model: comparison of clustering techniques in ambient and context-aware recommender systems | |
Chen et al. | A fuzzy matrix factor recommendation method with forgetting function and user features | |
CN103136309A (en) | Method for carrying out modeling on social intensity through learning based on core | |
Li et al. | Incorporating trust relation with PMF to enhance social network recommendation performance | |
Islek et al. | A hybrid recommendation system based on bidirectional encoder representations | |
He et al. | An empirical study on user-topic rating based collaborative filtering methods | |
Sobecki | Ant colony metaphor applied in user interface recommendation | |
Rathore et al. | Analysis on recommended system for web information retrieval using hmm | |
Guerraoui et al. | Sequences, items and latent links: Recommendation with consumed item packs | |
Zhang et al. | A CTR Prediction Model With Double Matrix-Level Cross-Features | |
Wang et al. | Cross-domain collaborative recommendation by transfer learning of heterogeneous feedbacks | |
Wang et al. | Collaborative filtering enhanced by user free-text reviews topic modelling |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |