Skip to Main Content
Kinds of topics and discussions come forth in Web forums every day, we can talk about newsletters and daily trivial matters in virtual communities, communicate with each other deeply in thought essentially. Part subjects are hot topics, which attract a lot of users, are widely viewed and massively discussed. Hot topic can be classified as isolated topic and social topic. The characteristics of social topic are multi-thematic, higher relevance between topics also. Isolated topics are lesser in quantity, a spot of relevance in it besides. Social topic' detecting algorithm is mainly based on subject relevance. This paper presents a density-based clustering model of subject words to detect social hot event from quantity and content relevance. Experiments on Tian YaZaTan community of TianYa BBS demonstrate the efficiency of the proposed model, extracting social topics which are better organized for search but also discovering communities of these topics.
Date of Conference: 11-13 Aug. 2012