标题：Chinese microblog topic detection based on the latent semantic analysis and structural property
作者：Yan, Xia ;Zhao, Hua
作者机构：[Yan, Xia ] Shenzhen Institute of Information Technology, Shenzhen, 518172, China;[Zhao, Hua ] College of Information Science and Engineering, Shandon 更多
来源：Journal of Networks
摘要：Traditional topic detection method can not be applied to the microblog topic detection directly, because the microblog text is a kind of the short, fractional and grass-roots text. In order to detect the hot topic in the microblog text effectively, we propose a microblog topic detection method based on the combination of the latent semantic analysis and the structural property. According to the dialogic property of the microblog, our proposed method firstly creates semantic space based on the replies to the thread, with the aim to solve the data sparseness problem; secondly, create the microblog model based on the latent semantic analysis; finally, propose a semantic computation method combined with the time information. We then adopt the agglomerative hierarchical clustering method as the microblog topic detection method. Experimental results show that our proposed methods improve the performances of the microblog topic detection greatly. © 2013 ACADEMY PUBLISHER.