Explore Dissimilarity Method for Summarization

This paper explores dissimilarity method for summarization. Understands the text based on significance of the sentences, and generates summarization based on extractive strategy. It computed the significance feature of the sentences, then computed the representative power of the sentences, and ordered them and selected the top sentences to generate the summary. The Chinese corpus and DUC corpus are used to experiment. Experimental result shows that the proposed summarization method can improve the performance of summarization.