A Summarization System with Categorization of Document Sets

We participated in both the single-document and multi-document summarization tasks at the TSC 2002. We have incorporated two modules into our earlier summarization system, which is based on a sentenceextraction technique, so that we could apply the system to the multi-document summarization task. One is a module to categorize document sets and the other is to estimate the similarity between sentences. The categorization of document sets is done based on extended named entity classes that include event or facility types as well as original classes. Our system uses the category information to decide how to use similarity information. The similarity between sentences is measured according to the Dice coefficient, and the results are used to either select a representative sentence from among similar sentences or to extract typical sentences from a given document set.