关键词: 话题演化, 子话题聚类, 内容特征, 时间特征

Abstract: Subtopic is the division for the topic, and it is a new research direction compared with the topic. Subtopic clustering is the base for the  analysis of topic evolution relations. This paper proposes a new method of clustering Chinese news subtopic integrating content and time features. It focuses on the analysis of subtopic content feature in text, and studies the computation of subtopic word weights and the dimension reduction of subtopic words. Five topics including 18 subtopics are used to conduct the experiment. Experimental results show that the performance of the proposed method is better than the existing subtopic identification methods.

Key words: topic evolution, subtopic clustering, content features, time features