Thread Specific Features are Helpful for Identifying Subjectivity Orientation of Online Forum Threads

Proceedings of COLING 2012: Technical Papers, pages 295–310, COLING 2012, Mumbai, December 2012. Thread Specific Features Are Helpful For Identifying Subjectivity Orientation of Online Forum Threads Prakhar Bi yani1 Sumit Bhatia2 Cornelia Caragea3 Prasenjit Mitra1 (1) College of Information Sciences and Technology, The Pennsylvania State University, USA (2) Department of Computer Science, The Pennsylvania State University, USA (3) Department of Computer Science and Engineering, University of North Texas, USA prakharbiyani@gmail.com, sumit@cse.psu.edu, ccaragea@unt.edu, pmitra@ist.psu.edu ABSTRACT Subjectivity analysis has been actively used in various applications such as opinion mining of customer reviews in online review sites, question-answering in CQA sites, multi-document summarization, etc. However, there has been very little focus on subjectivity analysis in the domain of online forums. Online forums contain huge amounts of user-generated data in the form of discussions between forum members on specific topics and are a valuable source of information. In this work, we perform subjectivity analysis of online forum threads. We model the task as a binary classification of threads in one of the two classes: subjective and non- subjective. Unlike previous works on subjectivity analysis, we use several non-lexical thread- specific features for identifying subjectivity orientation of threads. We evaluate our methods by comparing them with several state-of-the-art subjectivity analysis techniques. Experimental re- sults on two popular online forums demonstrate that our methods outperform strong baselines in most of the cases. KEYWORDS: Online Forums, subjecitivity, dialogue act. 295 1 Introduction A large number of online forums in various domains (e.g., health, sports, travel, camera, lap- tops, etc.) exists today, containing huge volumes of user-generated data in the form of discus- sions between members. The topics discussed in the threads of these forums are very unique in nature as they are often related to practical aspects of life (e.g., How much to tip after bad service?). Since such information is not available in other webpages, online forums are increasingly becoming very popular among internet users for discussing real life problems. In this work, we analyze subjectivity of online forum threads. We identify two types of threads in an online forum: subjective and non-subjective and we model the subjectivity analysis task as a binary classification problem. Subjective threads discuss subjective topics that seek opinions, viewpoints, evaluations, and other private states of people, whereas non-subjective threads discuss non-subjective topics that seek factual information. Figure 1 shows a subjective thread from an online forum, Trip–Advisor New York. Figure 2 shows a non-subjective thread from the same forum. In the former, the topic of discussion is whether to tip or not after bad service?, which seeks opinions, whereas the latter seeks factual information about bands/artists playing in December in Madison Square Gardens. To the best of our knowledge, previous work on subjectivity analysis has not tackled the problem of identifying subjectivity orientation of online threads. 1.1 Why Subjectivity Analysis of Online Forum Threads? • Improving forum search: Internet users search online forums, generally, for two types of information. Some of them search the forums for subjective information such as dif- …

Content truncated to 500 words