中文摘要：

为解决基于微博的用户兴趣建模存在建模不全面的问题，提出一种基于微博扩展的用户兴趣主题挖掘算法。通过结合用户自身兴趣及用户关注人的兴趣来扩充微博用户兴趣，将兴趣分为长期兴趣、过期兴趣、近期兴趣，利用改进的TF． IDF算法分析相关微博内容，利用基于时间及文档频率加权的主题词重要度计算兴趣主题词得分，得出综合全面的用户兴趣。实验结果表明，综合分析用户及用户关注人的兴趣，可以有效得到真实、全面、包含潜在兴趣的用户个人兴趣。

英文摘要：

To solve the problem that the modeling of modern user interest modeling based on Weibo is not comprehensive,a topic mining algorithm of user interest based on Weibo extension was presented which extended user interest by combing user interest with user’s follow friend interest.User interest was divided into user’s long-term interest,expire interest and recent interest. According to improved TF. IDF algorithm,user’s Weibo account was analyzed,improved subject headings importance calcula-tion algorithms based on time and document frequency weight was employed to analyze user interest,and then the comprehensive user interest was got.Experimental results show that,combining interests of user and user’s follow friends can get the user true personalization interest including potential information.