Using Query Expansion in Manifold Ranking for Query-Oriented Multi-document Summarization

Manifold ranking has been successfully applied in query-oriented multi-document summarization. It not only makes use of the relationships among sentences, but also between given query and sentences. However, information original is often insufficient. So we present a expansion method, which combined manifold to resolve this problem. Our method utilizes term itself knowledge base WordNet expand it by synonyms, uses document set various ways (mean expansion, variance TextRank expansion). Compared with previous methods, our combines multiple methods better represent information, at same time, useful attempt on ranking. In addition, degree word overlap proximity words calculate similarity We performed experiments datasets DUC 2006 DUC2007, evaluation results show that proposed can significantly improve system performance make comparable state-of-the-art systems.