Study On Topic Tracking And Tendency Classification Based On HowNet

Because the amount of information is too much, the information about topic is separated in different places and emerged in different time. Only by these isolated information, we can not grasp the event in whole. Topic Detection and Tracking (TDT) can help us to organize the potential information so that we can grasp all the details about events and the relations between events. Topic tracking involves tracking a given news event in a stream of news stories i.e. finding all subsequent stories in the news stream that discuss the given event.This paper focuses on the topic tracking task and topic tendency tracking, task, makes the research on the characteristic of topic and advanced topic tracking system of other research institutions. In view of the impact of transfer and differentiation of event on topic tracking, paper discusses the comprehensive methods of weight adjustment, event frame and story expansion to improve the tracking effectiveness. Finally paper proposes the algorithm of story tendency classification based on affective words and dynamic role frame in HowNet. The topic tendency of paper refers simple point of view that shows attitude to some event on third part. It falls into two kinds of attitude that is agree and oppose. It's a kind of information organization about topic. The research of topic tracking is an integration of knowledge expression , information retrieval and natural language processing which has become one of hot questions for discussion around world.Besides this, paper constructs the topic tracking system module based on war field and realizes the topic tracking and tendency tracking based on semantic resourse-HowNet. The experiment result shows that the research methods are effective.