Skip to Main Content
In this paper, a domain-specific extractive scheme combined with fuzzy logic and rough sets is proposed. In this scheme, the senses of domain-specific vocabulary constitute a fuzzy approximation space and a document is represented by a fuzzy set of senses. The use of word senses instead of the word itself is to cope with the problem that synonyms are treated differently. Also, a characteristic of the scheme is that the core concepts of the document are approximated by the lower approximation of its fuzzy set in the fuzzy approximation space. The results of applying domain-specific datasets to the scheme are shown. Their quality is measured by an intrinsic method.
Date of Conference: 5-7 Sept. 2007