Abstract
Nowadays, Netnews has become one of the most important channels for people to obtain information. The capacity for people to assimilate such vast amounts of information is limited. How to quickly and completely learn about the news of specific time turns into an urgent need. This paper proposes a method of bursty hot topic detection based on bursty feature. Firstly, determine the bursty features during a given time period based on term frequency, and cluster similar news stories together according to the content similarity and time decaying function to get the potential topic list, and then determine the final bursty hot topics based on bursty features. This method greatly reduces the complexity of the algorithm, improves efficiency, but also ensures the accuracy of bursty hot topic detection.