Identifying Multiple Userids of the Same Author

Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1124–1135, Seattle, Washington, USA, 18-21 October 2013. c⃝2013 Association for Computational Linguistics Identifying Multiple Userids of the Same Author Tieyun Qian Bing Liu State Key Laboratory of Software Eng. Wuhan University Department of Computer Science University of Illinois at Chicago 16 Luojiashan Road 851 South Morgan St., Chicago Wuhan, Hubei 430072, China IL, USA, 60607 qty@whu.edu.cn liub@cs.uic.edu Abstract This paper studies the problem of identifying users who use multiple userids to post in so- cial media. Since multiple userids may belong to the same author, it is hard to directly apply supervised learning to solve the problem. This paper proposes a new method, which still uses supervised learning but does not require train- ing documents from the involved userids. In- stead, it uses documents from other userids for classifier building. The classifier can be applied to documents of the involved userids. This is possible because we transform the document space to a similarity space and learning is performed in this new space. Our evaluation is done in the online review do- main. The experimental results using a large number of userids and their reviews show that the proposed method is highly effective. 1 Introduction It is common knowledge that some users in social media register multiple accounts/userids to post articles, blogs, reviews, etc. There are many rea- sons for doing this. For example, due to past post- ings, a user may become despised by others. He/she then registers another userid in order to regain his/her status. A user may also use multiple userids to instigate controversy or debates to popu- larize a topic to make it “hot” or even just to pro- mote activities at a website. Yet, a user may also use multiple userids to post fake or deceptive opin- ions to promote or demote some products (Liu, 2012). It is thus important to develop technologies to identify such multi-id users. This paper deals with this problem based on writing style and other linguistic clues. Problem definition: Given a set of userids ID = {id1, …, idn} and each idi has a set of documents Di, we want to identify userids that belong to the same physical author. The main related works to ours are in the area of authorship attribution (AA), which aims to identify authors of documents. AA is often solved using supervised learning. Let A = {a1, …, ak} be a set of authors (or classes) and each author ai  A has a set of training documents Di. A classifier is then built to decide the author a of each test document d, where a  A. We will discuss this and other re- lated works in Section 2. This supervised AA formulation, however, is not suitable for our task because we only have userids but not real authors. Since some of the userids may belong to the same author, we cannot treat each userid as a class because in that case, …

Content truncated to 500 words