A recommendation model based on collaborative filtering and factorization machines for social networks

The communication of persons based on mobile Internet is more and more frequency as the widespread of social network media like Facebook, Weibo and so on. Faced with the vast amount of user groups and variety user features, the traditional single recommended algorithm turns to noneffective. In this paper, we use the collaborative filtering to select features in social network relationship which is used to build a vector space model. Also, we propose a distributed algorithm based on factorization machines and genetic algorithm, and use the vector space model as one part of input. We split the records into different sessions depend on the time sequence for the filter of train dataset. In the experiment part we proved that all the methods mentioned above remarkable improved the recommendation accuracy.