Query-focused multidocument summarization based on hybrid relevance analysis and surface feature salience

摘要: Query-focused multidocument summarization is to synthesize from a set of topic-related documents brief, well-organized, fluent summary for the purpose answering an information need that cannot be met by just stating name, date, quantity, etc. In this paper, task essentially treated as sentence retrieval task. We propose hybrid relevance analysis evaluate query. This achieved combining similarities computed vector space model and latent semantic analysis. Surface features are also examined discern impact low-level query-focused summarization. addition, modified Maximal Marginal Relevance proposed reduce redundancy taking into account shallow feature salience. The experimental results show method obtained competitive when evaluated with DUC 2005 corpus.