Extraction of topic transition from document stream based on hierarchical clustering
暂无分享,去创建一个
Abstract We propose a method for extracting keywords expressing topic transition from document stream such as news articles based on hierarchical clustering and C-value method for constructing compound words. Through the user evaluation of 640 topics extracted between 32 days, we found users could understand 94.3% topics as news, and 68.6% topics including topic transition.
[1] Hiroshi Nakagawa,et al. A Simple but Powerful Automatic Term Extraction Method , 2002, COLING 2002.
[2] Sophia Ananiadou,et al. Extracting Nested Collocations , 1996, COLING.