User’s Behavior Analysis Based On Data Mining

With the rapid development of Internet, the Weibo has become an import platform for information sharing. Study on user’s behavior can provide theoretical basis for Weibo popularizing and monitoring.Based on SinaWeibo, the most popular weibo in China, this paper do many research on SinaWeibo data collecting and user clustering. Detailed tasks mainly include:1. Weibo information collecting system. This paper developed a SinaWeibo information collecting system. The system gets user and message information using web analysis.2. Abnormal user analysis. This paper analyzed behavior of abnormal user. First we transfer user’s attribute into words and generate it into attribute document, then we use LDA topic model for clustering and analysis.3. User interest topic analysis. This paper analyzed topic that user concerns about. First we participle user’s message and extract the keywords, then we use LDA topic model clustering user’s interested topic.