基于两层聚类的微博热点话题发现算法研究

Research on Hot Topic Discovery Algorithm of Microblog Based on Two-level Clustering

微博热点话题发现是指从大量的微博文本中发现用户讨论的热点话题,话题发现主要通过文本聚类的方法实现,聚类算法的选择和改进通常对结果有着重要的影响。针对微博话题发现任务,论文提出通过改进的SinglePass算法和层次聚类的方法,完成微博的话题发现,并且根据横向和纵向对比分析,验证算法话题发现的有效性。

Microblog hot topic discovery refers to the discovery of hot topics discussed by users from a large number of microblog texts.Topic discovery is mainly realized by text clustering.The selection and improvement of clustering algorithm usually have an important impact on the results.Aiming at the task of topic discovery in microblog,this paper proposes an improved single pass algorithm and hierarchical clustering method to complete topic discovery in microblog,and verifies the effectiveness of the algorithm based on the horizontal and vertical comparative analysis.