Sentence Selection for Extractive Summaries in PARE

1. ABSTRACT As the amount of text available from electronic sources continues to increase, the study of automatic text summarization is more important now than it has ever been. Automatically generated summaries should be clear and concise without giving unimportant or redundant information. The process of extractive text summarization involves determining the most important sentences in a document and putting them together to make a summary. We have developed and tested methods for determining which sentences should be included in single-document summaries generated by PARE, an automatic text summarizer. Evaluation of our results using ROUGE, an automated system for comparing summaries, showed that summaries generated by our sentence selection method compare favorably to commercially available text summarizers.