QueSTS: A Query Specific Text Summarization System

Effective extraction of query relevant information present within documents on the web is a nontrivial task. In this paper we present our system called QueSTS, which does the above task by filtering and aggregating important query relevant sentences distributed across a set of documents. Our approach captures the contextual relationships among sentences of all input documents and represents them as an "integrated graph". These relationships are exploited and several subgraphs of integrated graph which consist of sentences that are highly relevant to the query and that are highly related to each other are constructed. These subgraphs are ranked by our scoring model. The highest ranked subgraph which is rich in query relevant information and also has sentences that are highly coherent is returned as a query specific summary.

• Type: Conference
• Year: 2008
• Where: FLAIRS
• Authors: M. Sravanthi, C. Ravindranath Chowdary, P. Sreenivasa Kumar
• Keywords: Artificial Intelligence, FLAIRS 2008, Query Relevant Information, Query Relevant Sentences, Relevant Information Present