Modeling Mutual Influence Between Social Actions and Social Ties

Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers, pages 848–859, Dublin, Ireland, August 23-29 2014. Modeling Mutual Influence Between Social Actions and Social Ties Xiaofeng YU Junqing XIE HP Labs China Universal Business Park 10 Jiu XianQiao Road, Chaoyang District, Beijing, China {xiaofeng.yu,jun-qing.xie}@hp.com Abstract In online social media, social action prediction and social tie discovery are two fundamental tasks for social network analysis. Traditionally, they were considered as separate tasks and solved inde- pendently. In this paper, we investigate the high correlation and mutual influence between social actions (i.e. user-behavior interactions) and social ties (i.e. user-user connections). We propose a unified coherent framework, namely mutual latent random graphs (MLRGs), to flexibly encode evidences from both social actions and social ties. We introduce latent, or hidden factors and coupled models with users, users’ behaviors and users’ relations to exploit mutual influence and mutual benefits between social actions and social ties. We propose a gradient based optimization algorithm to efficiently learn the model parameters. Experimental results show the validity and competitiveness of our model, compared to several state-of-the-art alternative models. 1 Introduction With the dramatically rapid growth and great success of many large-scale online social networking ser- vices, social media bridge our daily physical life and the virtual Web space. Popular social media sites (e.g., Facebook and Twitter) and mobile social networks (e.g., Foursquare) have gathered billions of act- ing users and are still attracting millions of newbies everyday. Modeling social actions and social ties are two fundamental tasks in online social media. Social actions are the users’ activities or behaviors in socially connected networks. For example, a social action can be “posting a tweet” on Twitter or the “check-in” behavior on Foursquare. A social tie or social relation is referred to any relationship between two or more individual users in a social network, such as the friend and colleague relationships. By understanding a user’s behaviors and accordingly exploiting potentially interesting services to her/him, one can improve the user’s experience and boost the revenue of social media sites. Also, precise social tie prediction will help people tap into the wisdom of crowds, to aid in making more informed decisions. Since individual users are socially connected, social influence occurs through information diffusion in social networks. Social influence happens when one’s opinions or behaviors are affected by others. It is well known that different types of social ties have essentially different influence on social actions. Intuitively, a user’s trusted friends on the web affect that user’s online behavior. Ma et al. (2009) and Ma et al. (2011) claimed that one user’s final behavior decision is the balance between his/her own taste and her/his trusted friends’ favors. On the other hand, social actions also have important influence on social ties. Obviously, users with similar preferences or behaviors are more likely to be friends than others in social media. Users with momentous activities will attract many other users to be connected with. On the contrary, no body will be interested in …

Content truncated to 500 words