A comprehensive ranking model for tweets big data in online social network

Abstract

Online social network (OSN) is an important part of cyber physical system (CPS). In OSN, micro-blogging has grown rapidly to a popular online social network recently and provides a large number of real-time tweets for users. With the popularity of micro-blogging and the increase of active users, many users are faced with an information overload problem, especially for those with many followees and thousands of tweets arriving every day. In this paper, we aim to investigate the problem of recommending valuable tweets that users are really interested in personally, so as to reduce their efforts to find useful information. We consider three major aspects in our proposed ranking model, including the popularity of a tweet itself, the intimacy between the user and the tweet publisher, and the interest fields of the user. The detailed indicators for each aspect are introduced by analyzing users’ behaviors and their meanings on micro-blogs. The experimental results show that the proposed model can help improve the ranking performance in precision and greatly outperform several baseline methods.