An Algorithm of Chinese Micro-Blog Hot Topic Detecting Based on Clustering

Author:
Publisher website
Google Scholar
Abstract
Nowadays Micro blog has become a new tool of social intercourse. It effectively integrates various multimedia so that it is possible for users to release and update information freely. The characteristics of the micro blog bring great challenges to hot topic detecting. This paper proposed an approach of mining hot topics from micro blog by detecting the key terms emerging in large numbers and clustering them. To extract subject headings, a compound weight was structured by considering the term frequency and the growth rate to measure the likelihood of a word to be a hot topic keyword; To construct the hot topic, contextual relevance model was used to support incremental clustering, which is more suitable to the problem compared with semantic similarity. The experiments on real world data in micro blog show the effectiveness of the approach out of massive messages.