Extractive summarization using a fuzzy-rough-based relevance measure

In this paper, a novel method is proposed to extract key sentences of a document as its summary by estimating the relevance of sentences through the use of fuzzyrough sets. This method uses senses rather than raw words to lessen the problem that sentences of the same or similar semantic meaning but written in synonyms are treated differently. Also included is semantic clustering, used to avoid selecting redundant key sentences. A prototype of this automatic text summarization scheme is constructed and an intrinsic method with criteria widely used in information-retrieval systems is employed for measuring the summary quality. The results of applying the prototype to a dataset with manually-generated summaries are shown.