An exploration of document impact on graph-based multi-document summarization

摘要: The graph-based ranking algorithm has been recently exploited for multi-document summarization by making only use of the sentence-to-sentence relationships in documents, under assumption that all sentences are indistinguishable. However, given a document set to be summarized, different documents usually not equally important, and moreover, specific differently important. This paper aims explore impact on performance. We propose document-based graph model incorporate document-level information sentence-to-document relationship into process. Various methods employed evaluate two factors. Experimental results DUC2001 DUC2002 datasets demonstrate good effectiveness proposed model. Moreover, show robustness