Recommendation on Social Network Based on Graph Model

Twitter social network site provides a powerful means of sharing, organizing and finding content and contacts. The social network web of twitter forms a large graph; whose vertices are people and edges are relationships of the person. Twitter social networking is a typical of complex networks. Understanding the complex network is important with studying the characteristics of twitter network at large scale. Our interests in twitter focus on its complex network properties, such as scale free effect and small world effect. Here we demonstrate that the twitter social network is a scale free network and a small world network. A good recommender system is important for the social network web. The properties of the twitter network graph provide a theoretical basis for recommendation. In this work, we propose a graph-based recommendation algorithm using the relationship of users and adopt the Random Walk with Restarts to generate the recommendation users and evaluate the performance over precision-recall graph. The results show that recommendation based on the graph model performs well benefits from the relationship.