Research And Realization Of Hot Topic Discovery System Based On Sina Weibo

With the web2.0era, microblogging’s rapid development, has made it play an important role on the birth and transmission of hotspot. How to find hotspots from the vast amounts of microblogging data becomes an urgent demand of internet information.Based on the hot topic detection research, it was introduced in this thesis the main technology. And under the inspiration of the traditional text clustering, combined with the characteristics of microblogging, the idea that using two-phase clustering methods and hot topic sorting and show was proposed, on which a microblogging’s hot topic detection system was then designed. The system firstly discussed how to crawl microblogging data with the combination sina microblogging’s open API and dynamic web analysis technology, then described the process of hot topic detection, after improving the traditional K-Means and BIRCH clustering method. Considering user behavior information, the sorting and showing of hot topic were included. Finally, the validity and accuracy of the system was verified through the analysis of the test results.