A Hybrid On-line Topic Groups Mining Platform

Cheng-Lin Yang, Yun-Heh Chen-Burger

Abstract: In recent years, there is a rapid increased use of social networking platforms in the forms of short-text communication. Such communication can be indicative to popular public opinions and may be influential to real-life events. It is worth to identify topic groups from it automatically so it can help the analyst to understand the social network easily. However, due to the short-length of the texts used, the precise meaning and context of such texts are often ambiguous. In this paper, we proposed a hybrid framework, which adapts and extends the text clustering technique that uses Wikipedia as background knowledge. Based on this method, we are able to achieve higher level of precision in identifying the group of messages that has the similar topic.