Forward or ignore: user behavior analysis and prediction on microblogging

  • Authors:

  • Guanghua Song;Zhitang Li;Hao Tu

  • Affiliations:

  • College of Computer Science and Technology of HUST, Wuhan, China;College of Computer Science and Technology of HUST, Wuhan, China,Network Center, Huazhong University of Science and Technology, Wuhan, China;College of Computer Science and Technology of HUST, Wuhan, China,Network Center, Huazhong University of Science and Technology, Wuhan, China

  • Venue:
  • IEA/AIE'12 Proceedings of the 25th international conference on Industrial Engineering and Other Applications of Applied Intelligent Systems: advanced research in applied artificial intelligence
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

There has been an enormous development in online social networks all over the world in current times. Represented by Twitter and Facebook, the wave of online social networking is bringing broad impact and changing people's lives increasingly. At the same time, the online social networks are experiencing a rapid development in china. Large numbers of Chinese Internet users are spending more and more time on online social networks. Represented by SINA Weibo, the online social networks are gradually occupying Chinese people's vision and causing widespread concern. At present, the study of online social networks has focused on Twitter and Facebook, the popular Chinese online social network SINA Weibo has not been deeply studied. In this paper, we analyze the user's behavior on the SINA Weibo, pointing out the impact of user behavior in four key factors: the user's authority, the user's activity, the user's preferences and the user's social relations. By empirical methods, we give each factor the impact of user behavior through the likelihood. We find that the user's preferences and activity have greater impact on user behavior, while the authority of the user's social relations and values ​​of the user's behavior also has some impact. On this basis, we present an idea with machine learning to predict the behavior of users, and use pattern classification methods to solve the prediction problem. To the best of our knowledge this work is the first quantitative study on user behavior analysis. Changing the prediction problem into a pattern classification problem is the most important contribution of our work.