Research On Chinese Microblogging Topic Detection Based On User Authority

With the rapidly development of the Internet technologies, Networking Media has become one of the indispensable media. Specially, Weibo as a typical representative media has affected pattern of social communication with extremely speed. Users can update their status and discuss topics concerned anytime anywhere, which has made Weibo the center gathering hot topic. Searching the topics in Weibo, not only can it offers hot topic information, but also can support government to detect and analyze emergencies. Therefore, researching with respect to extract hot topic from massive amount of information has significant meaning in reality.Weibo text relative to traditional text is quite different, in addition to a large number of omissions and demonstrative pronoun, subjective personalized language, as well as the length of the text, the topic discrete, real-time and interactive features. Therefore, the traditional text topic searching methods cannot be directly applied to Weibo. This paper proposes authority based searching method towards Weibo topic by combining the features of Weibo.First of all, the searching algorithm introducing authority by taking into account of the number of fans as an important factor compares with searching algorithm only towards non-structured text, which improves the searching accuracy. Secondly, the traditional searching algorithm does not concern the time, nevertheless, the algorithm this paper proposed treats time factor as an essential parameter to make searching result more accurate by limiting the time. At last, limitation of traditional searching algorithm is an inability to sort the result or to sort the result directly. The scheme in this paper is able to sort the result by introducing the concept of fundamental energy and relative energy that can be directly applied.