Fast and accurate identification of implicit enterprise users in social media

There are a large number of implicit enterprise users in social media, who register as ordinary users but act like enterprise ones. It is a fundamental task to eliminate this type of users from the user set since their existing will severely affect the performance of many applications like network demographic investigation and targeted advertisement. Despite of its importance, this problem is surprisingly unexplored. In this paper, we present a novel two-stage framework to identify the implicit enterprise users. In the first stage, we detect this type of users by only using the profile information. In the second stage, we improve the classification accuracy by integrating more information from the user's contents. Such a framework meets well the scenario in reality: 1) due to the very nature of preprocessing, the identification of implicit enterprise users needs to be handled quickly with little cost, and 2) if enough resources such as microblog data and training/test time are provided, the problem is expected to be solved more precisely. We conduct extensive experiments on a real data set consisting of 400 users (200 ordinary users and 200 implicit enterprise users, respectively) in Sina Weibo, one of the largest social media in China. The result reaches an accuracy of 81.75% when using only profile features and is further improved to 86.00% when the content features are combined.