Research on Micro-Blog Information Perception and Mining Platform

To predict the tendency of Micro-blog information dissemination, provide the early warning of the Internet emergencies, and contribute to the content security of micro-blog, the paper offers a platform for Micro-blog information perceiving and mining. This platform is an integration of Micro-blog data collection and processing module, topic detection and tracking module, user behavior analysis module, trend prediction module, etc. It could access and analyze micro-blog information automatically, leading a positive significance to grasp the emergencies on micro-blog. This paper puts forward methods based on the Latent Dirichlet Allocation (LDA) document clustering and hot topics prediction, which could analysis and predict the micro-blog data effectively, avoiding the problems in the traditional algorithm. Also, these methods have a higher accuracy for clustering and prediction.