Identifying Information Units for Multiple Document Summarization

Multiple document summarization is becoming increasingly important as a way of reducing information overload, particularly in the context of the proliferation of similar accounts of events that are available on the Web. Removal of similar sentences often results in either partial or unwanted elimination of important information. In this paper, we present an approach to split sentences into their component clauses and use these clauses to produce comprehensive summaries of multiple documents describing particular events. Detailed analysis of all clauses and clause boundaries may be complex and computationally expensive. Our rule-based approach demonstrates that it is possible to achieve high accuracy in reasonable time.