Hot Topics Detection For Chinese Microblog

In recent years,microblogging has become a very important typical application of social network.People can release what they see,what they hear and what they think through the microblogging in any place and at any time.The more liberal approach of releaseing news greatly reduces the barriers to information dessemination,which is one of most important factor for widely using.In these microblogging data,majority of the microblogging information and real life are associated.some mocroblog websites list the keywords in recent period time in the home page.but these weibo keywords are too brief and are isolated in each period. Users can not have a comperehensive knowledege of the event thouguh these keywords. because cannot fully grasp the situation,The relevant departments can not properly guide public opinion.in the case,Hot topics detection for Chinese microblog gradually become a research hotspot.The main work of hot topics detection for chinese microblog is to organize the dispersed information and present through the relevant topic detection technology.This paper puts forward a complete weibo hot topic discovery method. The method includes weibo data text representation, feature extraction, text clustering and event keyword extraction steps, etc. In the paper,characteristics of weibo data itself are analysed carefully.and In the back of the data processing,we combined with the feature of the microblog data. At the same time,though improving the traditional event detection methods, The effect of the microblogging hot topics found is approved.we adopt multiple attribute undirected weighted graph clustering algorithm,not only consider the topology structure of figure but also introduce the node properties, Compared with common clustering algorithms to improve the clustering effect. Given some of the experimental results show that the proposed weibo hot event detection algorithm is effective.