Social Network Recommendation Based-on User Context

With the development of Internet technology, as a kind of informationdetection technology, recommendation has been getting more and moreattention of people. Recommendation technology of the social network hasalso aroused people’s concern. The main reason is that the development ofWeb2.0makes it more and more easy to create and share information,which causes the information explosion and the increasingly difficulty ofinformation discovery. In social network, compared with traditional targetedsearch, recommendation technology is the way to find information more inline with the information of human habits.The traditional recommendation technologies are mainly based oncollaborative filtering recommendation. While in social network, thecollaborative filtering method can no longer be used to fully describe thecomplex relations of social networks. So in order to solve the problem of therecommendation in social networks, we have to solve two problems. First isthe behavior model of how users behaving in a social network. This problemis a sociological problem, that means we have to find factors which affectthe user’s behavior and puts forward a realistic user behavior model whichreflects the various factors in the social network influencing each other andthe impact on user behavior. Second is to work out a good recommendationalgorithm. The algorithm needs to reflect the user behavior model, andconducts a more precise recommendation result of using the existing dugdata.To solve above problems, this paper first analyzes the characteristics ofsocial network and social network user behavior characteristics and then puts forward a user behavior model which considers both the social relationsin the social network, user personalization preferences and also thereal-time dynamic situational factors when user is browsing the network.Then combined with the existing recommendation approaches, thebehavioral model is formulized and a three-dimensional network ofrecommendation algorithm based on a user context is proposed, in whichthe influence of the possibility of eventually getting users to adopt theinformation is conducted by calculating the dynamic situations users ofinformation content and information to the amount of the sender to the user.Based on the above method, this paper has carried on the socialnetwork recommended method based on the user context. Then, experimentand analysis works have been conducted to do the verification of the designand implementation recommended method using the crawled Sinamicro-blog data. Experimental results show that the solution canrecommend more accurate content to the user more accord with the taste ofthe user in particular social network contexts.