CN104915391A - Article recommendation method based on trust relationship - Google Patents
Article recommendation method based on trust relationship Download PDFInfo
- Publication number
- CN104915391A CN104915391A CN201510272410.6A CN201510272410A CN104915391A CN 104915391 A CN104915391 A CN 104915391A CN 201510272410 A CN201510272410 A CN 201510272410A CN 104915391 A CN104915391 A CN 104915391A
- Authority
- CN
- China
- Prior art keywords
- user
- model
- trust
- article
- recommended
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an article recommendation method based on a trust relationship. The method includes the following steps that a user model and an article model are established, an old user-article scoring model and a user-user trust model are established, and recommended users are determined; according to the user-user trust model, an initial trust model is defined as direct trust; for old users, the similarity between the recommended users and the other old users is calculated according to the old user-article scoring model, and the most similar user set of the recommended users is obtained in an integration mode through the trust relationship of the recommended users; for new users, the most similar user set of the recommended users is calculated through the trust relationship between the new users and the old users; article scores are predicted on the most similar user set through a collaborative filtering algorithm, and high-scored articles are recommended. The method can solve the data sparsity problem in an existing recommendation system and the cold start problem of the users.
Description
Technical field
The invention belongs to artificial intelligence field, be specifically related to a kind of item recommendation method based on trusting relationship.
Background technology
Along with the development of infotech and internet, the various data such as user profile, merchandise news exponentially increase, occur " information overload ", be that information consumer or information producer encounter very large challenge: as information consumer, from bulk information, how find oneself interested information to be a very difficult thing; As information producer, how allowing the information of oneself producing show one's talent, receive the concern of users, is also a very difficult thing.Recommend method is exactly the important tool solving this contradiction.
At present, nearly all big-and-middle-sized website, such as Facebook, Amazon, Netflix, Taobao, bean cotyledon net etc. all in various degree employ various forms of recommend method, and obtain significant effect.Recommend method has become a very important technology in ecommerce, social networks, also creates huge economic benefit.Recommend method is obtained for and develops on a large scale very much in theory and practice.
Emerged in large numbers various proposed algorithm at present, main method has content-based recommendation and collaborative filtering recommending.
Content-based recommendation is that historical behavior according to user (as browses record, evaluation etc. to project) construct personal interest collection of illustrative plates and document, calculate the similarity of recommended project and user interest document, by project recommendation high for similarity to user, such as Pandora music site, its musician and researchist have listened the song of head up to ten thousand from different singer in person, then to different qualities (the such as melody of song, rhythm, music and the lyrics etc.) mark, then, Pandora can calculate the similarity of song according to the gene of expert's mark, and to other music similar on gene of the music that user recommends and he likes before.Obviously, this mode depends on the accurate description to item attribute, and need the attribute information of a large amount of assembled items, performance is low, usually has the problem of cold start-up and extendability.
The method of another kind of main flow is collaborative filtering recommending.Be applied in the earliest in email solution Tapestry.Collaborative filtering is also divided into based on the collaborative filtering of user and the collaborative filtering based on project (article).Based on the article that the collaborative filtering of user recommends other users similar with his interest to like to user, the collaborative filtering based on article recommends the article similar with the article that he likes before to user.For example, collaborative filtering based on user only just need can calculate the similarity between user to the score information of article according to user in the past, and the thing then just those users high with targeted customer (recommended user) similarity can liked recommends targeted customer.In like manner, the collaborative filtering based on article can be calculated similarity between article according to article by the user liked, and then recommends.
Collaborative filtering well compensate for the deficiency of content-based proposed algorithm, by feat of self simple feature efficiently, all be widely used in a lot of field, but this method only relies on the history score information of user, do not utilize other data, itself also also exists the problem such as cold start-up, Sparse.
In recent years, domestic and international researchers are calculating based on trust and are carrying out large quantity research in personalized recommendation, mainly utilize trusting relationship to improve the performance of personalized recommendation, alleviate the Sparse sex chromosome mosaicism of recommendation to a certain extent, and achieve certain achievement in research.But existing method mainly pays close attention to calculating and the reasoning thereof of the explicit trusting relationship of user, a lot of valuable implicit trust relation is often left in the basket, and only considers the trusting relationship on current recommendation platform.
Therefore, a kind of recommend method more effectively solving cold start-up user and Sparse Problem is needed.
Summary of the invention
The present invention is directed to above-mentioned prior art Problems existing to make improvements, namely the technical problem to be solved in the present invention is to provide a kind of item recommendation method based on trusting relationship, can solve toward the Sparse sex chromosome mosaicism in commending system and cold start-up customer problem.
In order to solve the problems of the technologies described above, the invention provides following technical scheme:
Based on an item recommendation method for trusting relationship, comprise the steps:
S1, set up user model and object model, set up old user-article Rating Model, user-user trust model, determine recommended user;
S2, according to user-user trust model, initial trust model is defined as direct trust; If system definition user A directly trusts user B, user B directly trusts user C, and user A does not have direct trusting relationship in original state and user C, and so define user A and indirectly trust user C, this situation is 1 step transmission; 2 step transmission and 3 step transmission by that analogy, and in transmittance process, add trust decay; Calculate all trusting relationships of recommended user;
S3, for old user, calculate the similarity of recommended user and other old users according to old user-article Rating Model, in conjunction with the trusting relationship of recommended user, integrate the most similar users collection drawing recommended user; For new user, calculate the most similar users collection of recommended user according to trusting relationship between new user and old user; Then on most similar users collection, utilize collaborative filtering to predict article scoring, article of being marked by height are as recommendation.
The foundation of described user model: according to user's historical record, comprises the article of purchase and sets up user model to the contact between the scoring of article and user; For new user, contacting on system introducing internet between this user and other old users, sets up user model.
The invention has the beneficial effects as follows: the present invention is based on collaborative filtering, in conjunction with the trusting relationship between user history information and user, and Trust transitivity is carried out to trust network, find indirect trusting relationship potential between user, solve Sparse Problem.For cold start-up user, system, by importing user's contact of other websites, is carried out Trust transitivity, thus is realized the recommendation compared with high-accuracy.
Accompanying drawing explanation
Accompanying drawing is used to provide a further understanding of the present invention, and forms a part for instructions, together with embodiments of the present invention for explaining the present invention, is not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is Trust transitivity principle schematic;
Fig. 2 is the process flow diagram of the article commending system based on trusting relationship.
Embodiment
As shown in Figure 1-2, the present invention discloses a kind of item recommendation method based on trusting relationship, comprises the steps:
S1, set up user model and object model, set up old user-article Rating Model, user-user trust model, determine recommended user;
S2, according to user-user trust model, initial trust model is defined as direct trust; If system definition user A directly trusts user B, user B directly trusts user C, and user A does not have direct trusting relationship in original state and user C, and so define user A and indirectly trust user C, this situation is 1 step transmission; 2 step transmission and 3 step transmission by that analogy, and in transmittance process, add trust decay; Calculate all trusting relationships of recommended user;
S3, for old user, calculate the similarity of recommended user and other old users according to old user-article Rating Model, in conjunction with the trusting relationship of recommended user, integrate the most similar users collection drawing recommended user; For new user, calculate the most similar users collection of recommended user according to trusting relationship between new user and old user; Then on most similar users collection, utilize collaborative filtering to predict article scoring, article of being marked by height are as recommendation.
The present invention is directed to current use commending system widely, based on collaborative filtering, in conjunction with the trusting relationship between user history information and user, and Trust transitivity is carried out to strengthen the effect of recommendation to trust network.Describe the present invention below in conjunction with use embodiment.
Fig. 1 is the description to Trust transitivity theory of the present invention, each node on behalf user, and arrow represents direct trusting relationship, and such as user A directly trusts user B and user D, and user B directly trusts user C.Theoretical according to Trust transitivity, user A trusts user C, E indirectly by 1 step transmission; User F is indirectly trusted by 2 step transmission; User G is indirectly trusted by 3 step transmission.For the trusting relationship can transmitted through mulitpath, take the shortest path: such as user A to user C can pass through A-B-C, also can pass through A-D-E-C, according to shortest path principle, take A-B-C, be i.e. 1 step transmission.
And add on the basis of Trust transitivity and trust decay, suppose that initial direct trust value is t, trust attenuation parameter is θ, being then t* θ by the indirect trust values of 1 step transmission, is t* θ * θ by the indirect trust values of 2 step transmission, by that analogy.Transmit 2 steps in theory to 3 steps for best.
Fig. 2 is the process flow diagram of the commending system based on trusting relationship.Its recommend method comprises the steps:
Step 1, commending system starts;
Step 2, determines recommended user;
Step 3, judges recommended user whether history of existence record (such as having bought article, to article scoring etc.), if existed, enters step 4; If there is no, be new user, then enter step 6;
Step 4, imports user-article rating matrix (a kind of form of expression of user's historical record);
Step 5, the information imported according to step 4 calculates the similarity (can adopt Pearson similarity) of other users in recommended user and system;
Step 6, importing user-user and trust matrix (form of expression of the relation between a kind of user and user), for old user, is the trusting relationship in system between old user and old user; For new user, be integrate other platforms on new user and old user between trusting relationship;
Step 7, carries out Trust transitivity to the trusting relationship that step 6 imports according to Trust transitivity theory proposed above, and is kept in trusting relationship by result;
Step 8, calculates the trust value (comprise directly trust and indirectly trust) of recommended user to other users in system;
Step 9, the similarity that integration step 5 is calculated and the trust value that step 8 is calculated, comprehensively draw the degree of association of other users in recommended user and system; If recommended user is history of existence record not, then take steps the separately recommended user of trust value computing of 8 and the degree of association of system other users interior;
Step 10, according to the degree of association that step 9 is calculated, draws the proximal subscribers topMatches (namely with N number of user that recommended user-association degree is maximum) of recommended user;
Step 11, according to the proximal subscribers topMatches of step 10, utilizes collaborative filtering to calculate " the prediction scoring " of user to article;
Article are arranged by " prediction score value " by step 12 from high to low;
Step 13, using top n article as recommendation article;
Step 14, recommends to terminate.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, although with reference to previous embodiment to invention has been detailed description, for a person skilled in the art, it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein portion of techniques feature.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (2)
1. based on an item recommendation method for trusting relationship, it is characterized in that, comprise the steps:
S1, set up user model and object model, set up old user-article Rating Model, user-user trust model, determine recommended user;
S2, according to user-user trust model, initial trust model is defined as direct trust; If system definition user A directly trusts user B, user B directly trusts user C, and user A does not have direct trusting relationship in original state and user C, and so define user A and indirectly trust user C, this situation is 1 step transmission; 2 step transmission and 3 step transmission by that analogy, and in transmittance process, add trust decay; Calculate all trusting relationships of recommended user;
S3, for old user, calculate the similarity of recommended user and other old users according to old user-article Rating Model, in conjunction with the trusting relationship of recommended user, calculate the most similar users collection of recommended user; For new user, calculate the most similar users collection of recommended user according to trusting relationship between new user and old user; Then on most similar users collection, utilize collaborative filtering to predict article scoring, article of being marked by height are as recommendation.
2. a kind of item recommendation method based on trusting relationship according to claim 1, is characterized in that, the foundation of described user model: according to user's historical record, comprises the article of purchase and sets up user model to the contact between the scoring of article and user; For new user, contacting on system introducing internet between this user and other old users, sets up user model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510272410.6A CN104915391A (en) | 2015-05-25 | 2015-05-25 | Article recommendation method based on trust relationship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510272410.6A CN104915391A (en) | 2015-05-25 | 2015-05-25 | Article recommendation method based on trust relationship |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104915391A true CN104915391A (en) | 2015-09-16 |
Family
ID=54084454
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510272410.6A Pending CN104915391A (en) | 2015-05-25 | 2015-05-25 | Article recommendation method based on trust relationship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104915391A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574430A (en) * | 2015-12-02 | 2016-05-11 | 西安邮电大学 | Novel privacy protection method in collaborative filtering recommendation system |
| 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 |
| CN109101667A (en) * | 2018-09-29 | 2018-12-28 | 新乡学院 | A kind of personalized recommendation method based on explicit trust and implicit trust |
| CN109690608A (en) * | 2016-02-29 | 2019-04-26 | Www.信任科学.Com股份有限公司 | The trend in score is trusted in extrapolation |
| CN111159570A (en) * | 2019-12-16 | 2020-05-15 | 聚好看科技股份有限公司 | Information recommendation method and server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100169328A1 (en) * | 2008-12-31 | 2010-07-01 | Strands, Inc. | Systems and methods for making recommendations using model-based collaborative filtering with user communities and items collections |
| CN102609523A (en) * | 2012-02-10 | 2012-07-25 | 上海视畅信息科技有限公司 | Collaborative filtering recommendation algorithm based on article sorting and user sorting |
| CN102779182A (en) * | 2012-07-02 | 2012-11-14 | 吉林大学 | Collaborative filtering recommendation method for integrating preference relationship and trust relationship |
-
2015
- 2015-05-25 CN CN201510272410.6A patent/CN104915391A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100169328A1 (en) * | 2008-12-31 | 2010-07-01 | Strands, Inc. | Systems and methods for making recommendations using model-based collaborative filtering with user communities and items collections |
| CN102609523A (en) * | 2012-02-10 | 2012-07-25 | 上海视畅信息科技有限公司 | Collaborative filtering recommendation algorithm based on article sorting and user sorting |
| CN102779182A (en) * | 2012-07-02 | 2012-11-14 | 吉林大学 | Collaborative filtering recommendation method for integrating preference relationship and trust relationship |
Non-Patent Citations (1)
| Title |
|---|
| 徐小伟: ""基于信任的协同过滤推荐算法在电子商务推荐系统的应用研究"", 《中国优秀硕士学位论文全文数据库 （电子期刊），信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574430A (en) * | 2015-12-02 | 2016-05-11 | 西安邮电大学 | Novel privacy protection method in collaborative filtering recommendation system |
| CN105574430B (en) * | 2015-12-02 | 2018-04-06 | 西安邮电大学 | A kind of new method for secret protection in Collaborative Filtering Recommendation System |
| CN109690608A (en) * | 2016-02-29 | 2019-04-26 | Www.信任科学.Com股份有限公司 | The trend in score is trusted in extrapolation |
| CN109690608B (en) * | 2016-02-29 | 2023-06-20 | Www.信任科学.Com股份有限公司 | Extrapolating trends in trust scores |
| 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 |
| CN109101667A (en) * | 2018-09-29 | 2018-12-28 | 新乡学院 | A kind of personalized recommendation method based on explicit trust and implicit trust |
| CN109101667B (en) * | 2018-09-29 | 2021-07-09 | 新乡学院 | Personalized recommendation method based on explicit trust and implicit trust |
| CN111159570A (en) * | 2019-12-16 | 2020-05-15 | 聚好看科技股份有限公司 | Information recommendation method and server |
| CN111159570B (en) * | 2019-12-16 | 2023-10-24 | 聚好看科技股份有限公司 | Information recommendation method and server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dhelim et al. | Personality-aware product recommendation system based on user interests mining and metapath discovery | |
| Yu | The dynamic competitive recommendation algorithm in social network services | |
| CN107784066A (en) | Information recommendation method, device, server and storage medium | |
| CN103886487A (en) | Individualized recommendation method and system based on distributed B2B platform | |
| CN104915391A (en) | Article recommendation method based on trust relationship | |
| CN105976229A (en) | Collaborative filtering algorithm based on user and project mixing | |
| CN104317835A (en) | New user recommendation method for video terminal | |
| CN104462383A (en) | Movie recommendation method based on feedback of users' various behaviors | |
| Kommineni et al. | Machine learning based efficient recommendation system for book selection using user based collaborative filtering algorithm | |
| Wu et al. | An approach of product usability evaluation based on Web mining in feature fatigue analysis | |
| CN113127754A (en) | Knowledge graph-based supplier recommendation method | |
| Zhenzhen et al. | Cross-domain item recommendation based on user similarity | |
| CN101957968A (en) | Online transaction service aggregation method based on Hadoop | |
| Stanescu et al. | A hybrid recommender system: User profiling from keywords and ratings | |
| Wang et al. | Personalized news recommendation based on consumers' click behavior | |
| CN106708871A (en) | Method and device for identifying social service characteristics user | |
| Xie et al. | A probabilistic recommendation method inspired by latent Dirichlet allocation model | |
| Margaris et al. | Improving collaborative filtering’s rating prediction accuracy by introducing the experiencing period criterion | |
| Cai et al. | Mechanism design for personalized recommender systems | |
| Kim et al. | Recommender systems using SVD with social network information | |
| Sun | Music Individualization Recommendation System Based on Big Data Analysis | |
| Pourgholamali | Mining information for the cold-item problem | |
| CN105528704A (en) | Dynamic product recommendation method based on stage | |
| Tang et al. | Service recommendation based on dynamic user portrait: an integrated approach | |
| Kumar et al. | Context-aware social popularity based recommender system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150916 |