Real-time topics extraction and visualization in online discussions

Abstract

Workshop line is a newly developed online discussion system. It is built on an enhanced version of chat line. The rules and other measures adopted in the workshop line discussions make it possible to extract topics with actual meanings in real time. In this paper, a tool for real-time topics extraction and visualization in workshop line discussions-THETA (a thesaurus-based topic analysis tool) is introduced. It is mainly based on a hierarchically structured thesaurus, which incorporates expertise. By simply scanning input text, counting frequencies of terms, summing up frequencies and sorting frequencies, topics with actual meanings can be extracted rapidly. Experimental results show that THETA works well and the real-time output can help online users to understand the contents of discussions and to discuss more fluently and efficiently.