Study on topic segmenting method in automatic abstracting system
Min Wan, Zhensheng Luo · 2004
Topic segmenting is an important problem in text structure analysis in automatic abstracting system. In order to get a higher quality abstract, we establish a vector space model based on simple semantic analysis for origin article, and then design an algorithm to find the borderlines between different topics by two sliding windows. Experimental results indicate that the method is feasible and useful.