CN102609533A - Kernel method-based collaborative filtering recommendation system and method - Google Patents

Kernel method-based collaborative filtering recommendation system and method Download PDF

Info

Publication number
CN102609533A
CN102609533A CN201210033951XA CN201210033951A CN102609533A CN 102609533 A CN102609533 A CN 102609533A CN 201210033951X A CN201210033951X A CN 201210033951XA CN 201210033951 A CN201210033951 A CN 201210033951A CN 102609533 A CN102609533 A CN 102609533A
Authority
CN
China
Prior art keywords
user
project
collaborative filtering
interest
similarity
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
Application number
CN201210033951XA
Other languages
Chinese (zh)
Other versions
CN102609533B (en
Inventor
俞能海
庄连生
王鹏
王晶晶
蒋锴
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology of China USTC
Original Assignee
University of Science and Technology of China USTC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Science and Technology of China USTC filed Critical University of Science and Technology of China USTC
Priority to CN201210033951.XA priority Critical patent/CN102609533B/en
Publication of CN102609533A publication Critical patent/CN102609533A/en
Application granted granted Critical
Publication of CN102609533B publication Critical patent/CN102609533B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention provides a kernel method-based collaborative filtering recommendation system and a kernel method-based collaborative filtering recommendation method. The corresponding system comprises a data preparation module which is used for standardizing the original data and carrying out corresponding preprocessing, generating a user-project rating matrix and a project distance matrix to output; a user interest modeling module which is used for constructing an interest model for a user on a project space according to the user-project rating matrix and the project distance matrix as well as a kernel density estimation technology; and a recommendation result generation module which is used for computing the similarities among the users according to the interest model, generating a neighbor set of a target user, and predicting a score of the project rated by the user according to a predetermined recommendation strategy and returning the recommendation result. Through the recommendation system and the recommendation method provided by the invention, the user interest model can be better presented, the user similarity in the practical application is estimated more accurately, the performance of the recommendation system can be promoted considerably, and more stable recommendation result can be obtained.

Description

A kind of collaborative filtering recommending system and method based on kernel method
Technical field
The invention belongs to that internet data excavates and information retrieval field, relate to a kind of commercial product recommending system and method in the ecommerce class website.
Background technology
Along with Information technology and WEB 2.0 technology rapid development, the huge day by day and maintenance rapid growth of internet information.For the Internet user, the problem that solve is how efficiently from magnanimity information, to excavate apace own valuable information; And for websites such as some social network sites, e-commerce websites, more will consider how effectively web site contents to be presented to the user, to improve service quality.The personalized recommendation technology progressively grows up under such background just; The main thought of this technology is the historical behavior record through digging user; Set up the interest model of describing user's request, utilize this interest model to go to find user and the direct relevance of information then, at last based on this relevance; With certain recommended models predict user preferences, thereby the information push that will meet this preference or demand supplies its selection to the user.In e-commerce website, system can perhaps browse record and user according to user's purchase the historical score data of commodity is recommended the interested product of its possibility; In multimedia sharing websites such as video sharing,, the user increased user's browsing time and the quantity that video is browsed greatly for recommending its interested video; At social network sites, good friend's recommendation has become it and has increased the user to the stickability of website and the important means of satisfaction.
A kind of as traditional recommended technology, the collaborative filtering recommending algorithm has also received a large amount of researchers' concern simultaneously because its advantages of simplicity and high efficiency characteristics have obtained more favor in practice.The collaborative filtering recommending algorithm can be divided into based on neighbours collection (Neighborhood-based) with based on two kinds of model (Model-based).Wherein, can be divided into again based on user (User-based) with based on the algorithm of project (Item-based) based on the proposed algorithm of neighbours collection.The core concept of collaborative filtering recommending algorithm is to utilize the approximate project of the project that approximate user or user like to filter bulk information, thereby filters out its possible interested project for the user.Particularly, based on user's collaborative filtering recommending algorithm will be similar with particular user interests the project recommendation that the user liked give this user; Project-based proposed algorithm then is to filter out the similar project of those projects of liking with the user as recommendation results.Collaborative filtering based on model is to utilize statistics and machine learning techniques to obtain a recommended models, and then produces recommendation results with this model.
Traditional collaborative filtering recommending algorithm need be considered ubiquitous drawback in the various collaborative filtering recommending algorithms; Mainly comprise: first; The data of common scoring have only been considered to have during the similarity of tradition method for measuring similarity between computational item or user; The user who causes only having the project of common scoring has similar possibility, is not inconsistent with actual conditions; Second; Collaborative recommendation is faced with the sparse challenge with the cold start-up problem of data; Under the sparse situation of score data, how can reasonably calculate similarity between the user, and then to produce the key issue that accurate recommendation results has become the quality that improves the collaborative filtering recommending algorithm.
Summary of the invention
The objective of the invention is to overcome the deficiency of traditional collaborative filtering recommending algorithm; A better user interest modeling scheme and corresponding with it similarity calculating method are provided; Thereby more accurately the digging user historical data promotes the raising of commending system performance.
The objective of the invention is to realize through following technical scheme:
A kind of collaborative filtering recommending system based on kernel method comprises:
Data preparation module is used for raw data standardization and corresponding pre-service are generated user-project rating matrix and project distance matrix and output;
The user interest MBM is used for according to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
The recommendation results generation module is used for according to said interest model, calculates the similarity between the user, generates neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
A kind of collaborative filtering recommending method based on kernel method comprises:
With raw data standardization and corresponding pre-service, generate user-project rating matrix and project distance matrix and output;
According to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
Based on said interest model, calculate the similitude between the user, generate neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
The present invention compared with prior art, its remarkable result is: not only considered the positive influences of data with existing to recommendation results during (1) digging user interest, also considered simultaneously the influence of negative scoring, can explain user interest model better; (2) user's similarity is no longer dependent on limited common scoring item, but fully excavates contact potential between project and the diffusion of user interest on the project space, has estimated the user's similarity in the practical application more accurately; (3) under the sparse situation of data, can promote lifting bigger on the commending system performance, and obtain more stable recommendation results.
Description of drawings
The structural representation that Fig. 1 provides for embodiment of the present invention based on the collaborative filtering recommending system of kernel method;
The distribution schematic diagram of user interest on the project space that Fig. 2 provides for embodiment of the present invention;
The schematic flow sheet that Fig. 3 provides for embodiment of the present invention based on the collaborative filtering recommending method of kernel method.
Embodiment
This embodiment provides a kind of collaborative filtering recommending system based on kernel method, and is as shown in Figure 1, comprising:
Data preparation module 11 is used for raw data standardization and corresponding pre-service are generated user-project rating matrix and project distance matrix and output;
User interest MBM 12 is used for according to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
Recommendation results generation module 13 is used for according to said interest model, calculates the similarity between the user, generates neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
Concrete, data preparation module 11 is responsible for preparing the needed data of total system, and partial data is carried out pre-service.System mainly uses two types of data: the user is to the historical score data of project and the category attribute information of project itself.Practical implementation is carried out as follows:
Steps A 1: initialization user-project rating matrix
User-project rating matrix is that the user representes the standardization of the historical score data of project, and its form is shown in form 1.The some users of each line display of this matrix are to the scoring of all items, and the scoring of all users to some projects shown in each tabulation.
Form 1
Figure BDA0000135925180000031
The value of each comprises two kinds of score data, R in the matrix M, nRepresent original score data; R * M, nThe historical score data of expression decentralization, the scope of promptly marking median is 0 situation.The former is used for marking at last forecast period, and the latter is used for the structure stage that the user builds model.If raw data is the data of decentralization, then both are identical.Allow the situation of scoring disappearance to exist here.
Steps A 2: the similarity of classifying between computational item
Next to carry out pre-service to the classified information of project itself.Suppose that the project category set for C, has comprised predefined several classifications among the C, the classification number is used | and C| representes.For project i, its category attribute is represented with the subclass of set C, is made as C i, C iComprise the one or more elements among the C.The formula of the classification similarity between computational item i and the project j is as follows:
Sim c ( i , j ) = | c i ∩ c j | 2 | c | * | c i ∪ c j |
This formula is weighed the similarity between two projects with one [0,1] interval rational number.To all items, calculate classification similarity between any two.The categorical attribute relative fixed of project, so but this step off-line carry out, and when having new projects to add, upgrade in time.
Steps A 3: the Pearson correlation coefficient between computational item
Similar with steps A 2, need calculate all items Pearson correlation coefficient between any two here.Pearson correlation coefficient computing formula between project i and the project j is as follows:
Corr ( i , j ) = Σ u ∈ U i , j ( R u , i - R i ‾ ) ( R u , j - R j ‾ ) Σ u ∈ U i , j ( R u , i - R i ‾ ) 2 Σ u ∈ U i , j ( R u , j - R j ‾ ) 2
U wherein I, jExpression all has the user of scoring to gather R to project i and project j U, iBe the value in the project rating matrix, represent the scoring of user u project i,
Figure BDA0000135925180000045
The average of the score value that is obtained for project i.
Steps A 4: the distance in the computational item space between projects
On the basis of steps A 2 and steps A 3, calculate the distance metric between projects, form the project distance matrix shown in form 2.
Form 2
Project 1 Project 2 ...... Project N
Project 1 D 1,1 D 1,2 ...... D 1,N
Project 2 D 2,1 D 2,2 ...... D 2,N
...... ...... ...... ...... ......
Project N D N,1 D N,2 ...... D N,N
D wherein N, NDistance between expression project N and the project N, computing method are as follows:
D i,j=1-Sim c(i,j)*Corr(i,j)
So far, the work of data preparation module is accomplished, and output user-project rating matrix and project distance matrix are as the input of subsequent module.
The method that user interest MBM 12 utilizes Density Estimator to user interest the distribution on whole project space estimate.Density Estimator is the nonparametric technique of density Estimation.In the method, kernel function and bandwidth thereof have multiple choices.This instructions is the practical implementation step of user interest modeling scheme among example explanation the present invention with the nucleus vestibularis triangularis function.
Step B1: calculate the kernel function value of user on all items
For a user, in existing score data, comprised the distribution of its interest.Suppose that it is I that user u comments undue project set uI uIt is the subclass of project set I.Before calculating, at first to preset the bandwidth h of kernel function, for example establish h=0.4.For project i, user's kernel function value is above that calculated as follows.To I uIn project j, calculate the influence of the scoring of j to project i:
If | D i , j | < 2 h , Then K i ( j ) = R u , j &times; ( 2 h - | D i , j | ) / 2 h ;
Otherwise, K i(j)=0;
Press as above method calculating kernel function value for all items, obtain the distribution plan of user interest on the project space as shown in Figure 2.
Step B2: the interest density of estimating user on all items
On the basis of steps A 1, estimating user-project interest density matrix is shown in form 3.
Form 3
Figure BDA0000135925180000053
The interest density of
Figure BDA0000135925180000054
expression user u on project i in the table, calculate according to following formula:
f ^ u ( i ) = 1 | I u | * h &Sigma; j &Element; I u K i ( j )
All users are repeated as above to calculate, up to whole user-project interest density matrix is filled up.So far, the task of user interest MBM is accomplished, and output user-project interest density matrix supplies the recommendation results generation module to use.
User-project rating matrix and user-project interest density Estimation matrix that recommendation results generation module 13 utilizes preceding two modules to produce are predicted the score data of disappearance, and recommendation results are returned to the user.Here so that targeted customer u is recommended as the workflow that example is introduced this module.Practical implementation is carried out according to four steps that are described below.
Step C1: the interest distribution similarity of calculating targeted customer and other users
In the user shown in the form 3-project interest density matrix; Each row can be regarded as a user interest model of simplifying version, and the interest distribution similarity of calculating user u and user v is equivalent to the similarity of the row vector that calculates two correspondences in user-project interest density matrix to a certain extent.In the specific embodiment of the invention, different with traditional similarity calculation method, weigh the similarity between the user interest distribution with the volume of two distribution laps.Particularly; For user u and user v, consider each project i with and on interest density
Figure BDA0000135925180000062
and
Figure BDA0000135925180000063
If f ^ u ( i ) * f ^ v ( i ) > 0 , Then Sim ( u , v ) = Sim ( u , v ) + Min ( | f ^ u ( i ) | , | f ^ v ( i ) | ) ,
Otherwise, Sim ( u , v ) = Sim ( u , v ) + Min ( f ^ u ( i ) , f ^ v ( i ) )
Just obtained the similarity of two user u and v after all items traversal one time.Repeat above calculating and can obtain the similarity between targeted customer u and other all users of data centralization.
Step C2: generate targeted customer neighbours' collection according to the interest distribution similarity
For targeted customer u, constitute its neighbours' collection with its similarity greater than those users of 0, similarity is that the user of negative value constitutes its negative neighbours' collection.
Step C3: prediction disappearance score data
For destination item i, target of prediction user u is shown below to its scoring:
P u , i = R &OverBar; u + &Sigma; v &Element; N ( u ) &cap; U i sim ( u , v ) * ( R v , i - R &OverBar; v ) &Sigma; v &Element; N ( u ) &cap; U i | sim ( u , v ) |
Wherein, neighbours' collection of N (u) expression user u, U iThe project i of being expressed as comments undue user's set.
Step C4: produce recommendation results and return to the targeted customer
For the project that the user does not mark, predict its scoring according to the method for step 3, then according to predicting that scoring order from high to low is to entry sorting.Based on the needs in the practical application, a preceding K project is returned to the user, K is preassigned value.
The technical scheme that adopts this embodiment to provide; Compared with prior art; Its remarkable result is: not only considered the positive influences of data with existing to recommendation results during (1) digging user interest, also considered the influence of negative scoring simultaneously, can explain user interest model better; (2) user's similarity is no longer dependent on limited common scoring item, but fully excavates contact potential between project and the diffusion of user interest on the project space, has estimated the user's similarity in the practical application more accurately; (3) under the sparse situation of data, can promote lifting bigger on the commending system performance, and obtain more stable recommendation results.
Embodiment of the present invention also provides a kind of collaborative filtering recommending method based on kernel method, and is as shown in Figure 3, comprising:
Step 31 with raw data standardization and corresponding pre-service, generates user-project rating matrix and project distance matrix and output;
Step 32, according to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
Step 33 according to said interest model, is calculated the similarity between the user, generates neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
Optional, corresponding pre-service comprises: said raw data all is converted into 0 to be the score data that comprises negative value of median, keeps original score data simultaneously.
Optional, corresponding project distance obtains through following method: through the distance between two projects on the distance metric function acquisition project space of classification similarity between project and Pearson correlation coefficient.
Optional, the user makes up through following method at the interest model on the project space: adopt the method for Density Estimator to set up user's the model of interest on the project space.
Optional, the similarity between the user obtains through following method: adopt probability distribution on two project spaces overlapping part estimate two users' similarity.
The implementation of the various method steps that comprises in the above-mentioned collaborative filtering recommending method based on kernel method is described in system's embodiment before, no longer is repeated in this description at this.
The technical scheme that adopts this embodiment to provide; Compared with prior art; Its remarkable result is: not only considered the positive influences of data with existing to recommendation results during (1) digging user interest, also considered the influence of negative scoring simultaneously, can explain user interest model better; (2) user's similarity is no longer dependent on limited common scoring item, but fully excavates contact potential between project and the diffusion of user interest on the project space, has estimated the user's similarity in the practical application more accurately; (3) under the sparse situation of data, can promote lifting bigger on the commending system performance, and obtain more stable recommendation results.
The above; Be merely the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, any technician who is familiar with the present technique field is in the technical scope that the present invention discloses; The variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.

Claims (10)

1. the collaborative filtering recommending system based on kernel method is characterized in that, comprising:
Data preparation module is used for raw data standardization and corresponding pre-service are generated user-project rating matrix and project distance matrix and output;
The user interest MBM is used for according to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
The recommendation results generation module is used for according to said interest model, calculates the similarity between the user, generates neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
2. the collaborative filtering recommending system based on kernel method according to claim 1 is characterized in that, in data preparation module, comprises:
The pre-service submodule is used for said raw data all is converted into 0 to be the score data that comprises negative value of median, keeps original score data simultaneously.
3. the collaborative filtering recommending system based on kernel method according to claim 1 is characterized in that, in the user interest MBM, comprises:
The distance calculation submodule is used for distance metric function through classification similarity between project and Pearson correlation coefficient and obtains on the project space distance between two projects.
4. the collaborative filtering recommending system based on kernel method according to claim 1 is characterized in that, in the user interest MBM, also comprises:
The modelling submodule is used to adopt the method for Density Estimator to set up user's the model of interest on the project space.
5. the collaborative filtering recommending system based on kernel method according to claim 1 is characterized in that, in the recommendation results generation module, comprises:
Similarity estimator module, be used to adopt two probability distribution on the project space overlapping part estimate two users' similarity.
6. the collaborative filtering recommending method based on kernel method is characterized in that, comprising:
With raw data standardization and corresponding pre-service, generate user-project rating matrix and project distance matrix and output;
According to said user-project rating matrix and project distance matrix, and through the interest model of Density Estimator technique construction user on the project space;
Based on said interest model, calculate the similitude between the user, generate neighbours' collection of targeted customer, and with predetermined recommendation strategy predictive user to the scoring of project and return recommendation results.
7. the collaborative filtering recommending method based on kernel method according to claim 6 is characterized in that, said pre-service comprises:
Said raw data all is converted into 0 to be the score data that comprises negative value of median, keeps original score data simultaneously.
8. the collaborative filtering recommending method based on kernel method according to claim 6 is characterized in that, said project distance obtains through following method:
Distance metric function through classification similarity between project and Pearson correlation coefficient obtains on the project space distance between two projects.
9. the collaborative filtering recommending method based on kernel method according to claim 6 is characterized in that, the interest model of said user on the project space makes up through following method:
Adopt the method for Density Estimator to set up user's the model of interest on the project space.
10. the collaborative filtering recommending method based on kernel method according to claim 6 is characterized in that, the similarity between the said user obtains through following method:
Adopt probability distribution on two project spaces overlapping part estimate two users' similarity.
CN201210033951.XA 2012-02-15 2012-02-15 Kernel method-based collaborative filtering recommendation system and method Expired - Fee Related CN102609533B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210033951.XA CN102609533B (en) 2012-02-15 2012-02-15 Kernel method-based collaborative filtering recommendation system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210033951.XA CN102609533B (en) 2012-02-15 2012-02-15 Kernel method-based collaborative filtering recommendation system and method

Publications (2)

Publication Number Publication Date
CN102609533A true CN102609533A (en) 2012-07-25
CN102609533B CN102609533B (en) 2015-03-18

Family

ID=46526905

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210033951.XA Expired - Fee Related CN102609533B (en) 2012-02-15 2012-02-15 Kernel method-based collaborative filtering recommendation system and method

Country Status (1)

Country Link
CN (1) CN102609533B (en)

Cited By (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103218407A (en) * 2013-03-22 2013-07-24 南京信通科技有限责任公司 Recommendation engine based on interest graph
CN103345503A (en) * 2013-07-01 2013-10-09 杭州万事利丝绸科技有限公司 Silk product individualized recommendation method based on wavelet network
CN103514304A (en) * 2013-10-29 2014-01-15 海南大学 Project recommendation method and device
CN104318268A (en) * 2014-11-11 2015-01-28 苏州晨川通信科技有限公司 Multiple transaction account identification method based on local distance measuring and learning
CN104731887A (en) * 2015-03-13 2015-06-24 东南大学 User similarity measuring method in collaborative filtering
CN104933595A (en) * 2015-05-22 2015-09-23 齐鲁工业大学 Collaborative filtering recommendation method based on Markov prediction model
CN106126727A (en) * 2016-07-01 2016-11-16 中国传媒大学 A kind of big data processing method of commending system
CN107105322A (en) * 2017-05-23 2017-08-29 深圳市鑫益嘉科技股份有限公司 A kind of multimedia intelligent pushes robot and method for pushing
CN107305559A (en) * 2016-04-21 2017-10-31 中国移动通信集团广东有限公司 Method and apparatus are recommended in one kind application
CN107818491A (en) * 2017-09-30 2018-03-20 平安科技(深圳)有限公司 Electronic installation, Products Show method and storage medium based on user's Internet data
WO2018090793A1 (en) * 2016-11-18 2018-05-24 腾讯科技(深圳)有限公司 Multimedia recommendation method and device
CN108182268A (en) * 2018-01-16 2018-06-19 浙江工商大学 A kind of collaborative filtering recommending method and system based on community network
CN108460145A (en) * 2018-03-15 2018-08-28 北京邮电大学 A kind of collaborative filtering recommending method based on mixing Interest Similarity
TWI643151B (en) * 2012-10-01 2018-12-01 美商菲絲博克公司 Method, computer-readable non- transitory storage medium, and system for predicting an interest of a user for content associated with another user
CN109495770A (en) * 2018-11-23 2019-03-19 武汉斗鱼网络科技有限公司 A kind of direct broadcasting room recommended method, device, equipment and medium
CN109636529A (en) * 2018-12-14 2019-04-16 苏州大学 A kind of Method of Commodity Recommendation, device and computer readable storage medium
CN109871215A (en) * 2017-12-05 2019-06-11 华为软件技术有限公司 The method and apparatus of software publication
CN111815410A (en) * 2020-07-07 2020-10-23 中国人民解放军军事科学院国防科技创新研究院 Commodity recommendation method based on selective neighborhood information
CN112667885A (en) * 2020-12-04 2021-04-16 四川长虹电器股份有限公司 Matrix decomposition collaborative filtering method and system for coupling social trust information
CN113852867A (en) * 2021-05-27 2021-12-28 天翼智慧家庭科技有限公司 Program recommendation method and device based on kernel density estimation

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020065797A1 (en) * 2000-11-30 2002-05-30 Wizsoft Ltd. System, method and computer program for automated collaborative filtering of user data
US20060282304A1 (en) * 2005-05-02 2006-12-14 Cnet Networks, Inc. System and method for an electronic product advisor
CN102129462A (en) * 2011-03-11 2011-07-20 北京航空航天大学 Method for optimizing collaborative filtering recommendation system by aggregation
CN102231166A (en) * 2011-07-12 2011-11-02 浙江大学 Collaborative recommendation method based on social context

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020065797A1 (en) * 2000-11-30 2002-05-30 Wizsoft Ltd. System, method and computer program for automated collaborative filtering of user data
US20060282304A1 (en) * 2005-05-02 2006-12-14 Cnet Networks, Inc. System and method for an electronic product advisor
CN102129462A (en) * 2011-03-11 2011-07-20 北京航空航天大学 Method for optimizing collaborative filtering recommendation system by aggregation
CN102231166A (en) * 2011-07-12 2011-11-02 浙江大学 Collaborative recommendation method based on social context

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
何绍华等: "基于核估计的电子商务协作过滤方法", 《计算机工程与应用》, no. 5, 20 February 2006 (2006-02-20), pages 207 - 209 *
康雨洁: "基于协同过滤的个性化社区推荐方法研究", 《中国优秀硕士学位论文全文数据库》, 20 September 2011 (2011-09-20), pages 11 - 19 *

Cited By (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI643151B (en) * 2012-10-01 2018-12-01 美商菲絲博克公司 Method, computer-readable non- transitory storage medium, and system for predicting an interest of a user for content associated with another user
US10257309B2 (en) 2012-10-01 2019-04-09 Facebook, Inc. Mobile device-related measures of affinity
CN103218407A (en) * 2013-03-22 2013-07-24 南京信通科技有限责任公司 Recommendation engine based on interest graph
CN103345503B (en) * 2013-07-01 2016-04-13 杭州万事利丝绸科技有限公司 A kind of silk product personalized recommendation method based on wavelet network
CN103345503A (en) * 2013-07-01 2013-10-09 杭州万事利丝绸科技有限公司 Silk product individualized recommendation method based on wavelet network
CN103514304B (en) * 2013-10-29 2017-01-18 海南大学 Project recommendation method and device
CN103514304A (en) * 2013-10-29 2014-01-15 海南大学 Project recommendation method and device
CN104318268A (en) * 2014-11-11 2015-01-28 苏州晨川通信科技有限公司 Multiple transaction account identification method based on local distance measuring and learning
CN104318268B (en) * 2014-11-11 2017-09-08 苏州晨川通信科技有限公司 A kind of many trading account recognition methods based on local distance metric learning
CN104731887A (en) * 2015-03-13 2015-06-24 东南大学 User similarity measuring method in collaborative filtering
CN104731887B (en) * 2015-03-13 2018-02-02 东南大学 A kind of user method for measuring similarity in collaborative filtering
CN104933595A (en) * 2015-05-22 2015-09-23 齐鲁工业大学 Collaborative filtering recommendation method based on Markov prediction model
CN107305559A (en) * 2016-04-21 2017-10-31 中国移动通信集团广东有限公司 Method and apparatus are recommended in one kind application
CN106126727A (en) * 2016-07-01 2016-11-16 中国传媒大学 A kind of big data processing method of commending system
WO2018090793A1 (en) * 2016-11-18 2018-05-24 腾讯科技(深圳)有限公司 Multimedia recommendation method and device
CN107105322A (en) * 2017-05-23 2017-08-29 深圳市鑫益嘉科技股份有限公司 A kind of multimedia intelligent pushes robot and method for pushing
CN107818491A (en) * 2017-09-30 2018-03-20 平安科技(深圳)有限公司 Electronic installation, Products Show method and storage medium based on user's Internet data
CN109871215A (en) * 2017-12-05 2019-06-11 华为软件技术有限公司 The method and apparatus of software publication
CN108182268B (en) * 2018-01-16 2021-01-08 浙江工商大学 Collaborative filtering recommendation method and system based on social network
CN108182268A (en) * 2018-01-16 2018-06-19 浙江工商大学 A kind of collaborative filtering recommending method and system based on community network
CN108460145A (en) * 2018-03-15 2018-08-28 北京邮电大学 A kind of collaborative filtering recommending method based on mixing Interest Similarity
CN108460145B (en) * 2018-03-15 2020-07-03 北京邮电大学 Collaborative filtering recommendation method based on mixed interest similarity
CN109495770A (en) * 2018-11-23 2019-03-19 武汉斗鱼网络科技有限公司 A kind of direct broadcasting room recommended method, device, equipment and medium
CN109636529A (en) * 2018-12-14 2019-04-16 苏州大学 A kind of Method of Commodity Recommendation, device and computer readable storage medium
CN109636529B (en) * 2018-12-14 2022-04-12 苏州大学 Commodity recommendation method and device and computer-readable storage medium
CN111815410A (en) * 2020-07-07 2020-10-23 中国人民解放军军事科学院国防科技创新研究院 Commodity recommendation method based on selective neighborhood information
CN111815410B (en) * 2020-07-07 2022-04-26 中国人民解放军军事科学院国防科技创新研究院 Commodity recommendation method based on selective neighborhood information
CN112667885A (en) * 2020-12-04 2021-04-16 四川长虹电器股份有限公司 Matrix decomposition collaborative filtering method and system for coupling social trust information
CN113852867A (en) * 2021-05-27 2021-12-28 天翼智慧家庭科技有限公司 Program recommendation method and device based on kernel density estimation

Also Published As

Publication number Publication date
CN102609533B (en) 2015-03-18

Similar Documents

Publication Publication Date Title
CN102609533B (en) Kernel method-based collaborative filtering recommendation system and method
CN102629257B (en) Commodity recommending method of e-commerce website based on keywords
CN110222272A (en) A kind of potential customers excavate and recommended method
CN101853463A (en) Collaborative filtering recommending method and system based on client characteristics
CN103886487A (en) Individualized recommendation method and system based on distributed B2B platform
CN110852852A (en) Industrial Internet product recommendation system and method
CN101482884A (en) Cooperation recommending system based on user predilection grade distribution
CN102968506A (en) Personalized collaborative filtering recommendation method based on extension characteristic vectors
CN104281956A (en) Dynamic recommendation method capable of adapting to user interest changes based on time information
CN105976229A (en) Collaborative filtering algorithm based on user and project mixing
CN103258020A (en) Recommending system and method combining SNS and search engine technology
CN103544632A (en) Method and system for individually recommending network commodities
CN105809558A (en) Social network based recommendation method and apparatus
CN104252679A (en) Construction method and system of brand advertisement evaluation system
CN103345699A (en) Personalized food recommendation method based on commodity forest system
CN110188268A (en) A kind of personalized recommendation method based on label and temporal information
CN104239496A (en) Collaborative filtering method based on integration of fuzzy weight similarity measurement and clustering
CN105023178B (en) A kind of electronic commerce recommending method based on ontology
CN101957968A (en) Online transaction service aggregation method based on Hadoop
CN105678590A (en) topN recommendation method for social network based on cloud model
CN104978406A (en) User behavior analysis method of Internet platform
CN102799656A (en) Target-user-oriented individualized product package extracting method
CN103744904A (en) Method and device for providing information
Sun et al. Leveraging friend and group information to improve social recommender system
Zhu et al. Consumer preference analysis based on text comments and ratings: A multi-attribute decision-making perspective

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20150318

Termination date: 20210215

CF01 Termination of patent right due to non-payment of annual fee