Behavior Analysis of Microblog Users Based on Transitions in Posting Activities

  • Authors:

  • Yamaguchi Yutaro;Shuhei Yamamoto;Tetsuji Satoh

  • Affiliations:

  • Graduate School of Library, Information and Media Studies, University of Tsukuba, 1-2 Kasuga, Tsukuba-city, Ibaraki, Japan;Graduate School of Library, Information and Media Studies, University of Tsukuba, 1-2 Kasuga, Tsukuba-city, Ibaraki, Japan;Graduate School of Library, Information and Media Studies, University of Tsukuba 1-2 Kasuga, Tsukuba-city, Ibaraki, Japan

  • Venue:
  • Proceedings of International Conference on Information Integration and Web-based Applications & Services
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

In recent years, such microblogs as Twitter have spread widely over the world. Twitter, which enables instant text communications among users, was launched in 2006. In 2012, its Japanese users exceeded 29.9 million. Useful functions related to posting a tweet include reply, retweet, and hashtag. Users communicate with others and spread information with these functions. In this paper, we model user behaviors by a transition of clusters that represent particular posting activities. Under the model, all users belong to a cluster consisting of several features at individual time slots and move among the clusters in a time series. These features include the number of posts and retweets/replies, the time when the tweets were posted, and the number of characters in each tweet. We reveal the temporal transitions of these clusters in the process of using Twitter from the time when users created their accounts. We propose a time longitudinal analysis method to clarify the relation of the transition of user posting activities and their lifetime of Twitter. Our proposed method consists of two steps: creating clusters that represent particular posting activities and drawing a state transition diagram with transition probabilities among clusters. From our analysis results of actual Japanese tweets for a one-year period with our proposed method, we conclude the following. Our proposed method can express changes in the posting activities of users. We conclude that using Twitter's functions, e.g., replies and retweets (RT), are one difference between users who continue to use Twitter for a long time and those who quit relatively soon.