Design Of Hot Event Detection System In Tibetan Web

After the advent of Internet media since1970s, we have stepped into an era of unprecedented wealth information, and the mode of transmission for information has undertaken great changes, more and more people are willing to express their views, ideas, and attitudes through network. As the information has not been organized and managed, This makes it more and more difficult to achieve our information,so it is an urgent need for a tool to obtain the information quickly.At present,internet users could achieve their information by search engine,because of using the keyword matching algorithm and therefore the results of high information redundancy. However, show us a lot of irrelevant things, so it would be difficult to grasp some new events fully. every year there will be news organizations selected in a field of hot events, but the time period in years, and the result is selected, the immediacy and objectivity of the results can not be guaranteed.The system designed by this paper based on the corpus of Tibetan report on the Internet, uses TDT technology to identify, track and cluster new events, designs a Hot Event Detection which hot events could be discovered by this system, during any period selected by users, and the results are objective.The corpus of the People’s Tibetan website, the use of TDT (Topic Detection and Tracking) technology to identify and track news events and news events clustering and then use the Web crawler to crawl the page within the specified range, extract the text and Tibetan segmentation to generateweighted vectors, propose a method for calculating event through researching algorithm, and improve the system’s sensitivity for new hot events, second-layer clustering strategy has been used to cluster the texts, at last, we get to the event list.Finally, we choose the news corpus to do experiments and certify the algorithms we mentioned above, conduct relevant evaluations, the result reveals that the system makes great achievements.