Query-focused multi-document summarization aims to create a compressed summary biased to a given query. This paper presents a context-sensitive approach based on manifold ranking of sentences to this summarization task. The proposed context enhanced manifold ranking approach not only looks at the sentence itself, but also considers its surrounding contextual information. Compared to the existing manifold ranking approach which totally ignores the contextual information of a sentence, this approach can capture more additional relevant information which is especially necessary for formulating the relationships between short text snippets like sentences. Experiments are conducted on the DUC 2005 and DUC 2006 data sets and the ROUGE evaluation results demonstrate the advantages of the proposed approach.
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||11th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2010|
|Country||Korea, Republic of|
|Period||30/08/10 → 2/09/10|
- Theoretical Computer Science
- Computer Science(all)