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

Abstract

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