Knowledge Acquisition Method Research Based On Fuzzy Rough Set

In order to deal with the knowledge of incomplete, uncertainty, and inharmonious in the real word, Poland scholar Pawlak proposed rough set concepts and theory in1982. It discussed the reduction theory on the approximate space which constituted by universe and the equivalence relation on the universe. However this model often had great limitations in the practical problems, therefore, many scholars extended the Pawlak rough sets. Among them, fuzzy rough set combined with the rough set and fuzzy set can be more effective to deal with knowledge base which contains continuous and fuzziness value.The knowledge acquisition method based on the fuzzy rough set played a very important role in the research of fuzzy rough sets. The matrix characterizations of concepts in rough set theory are proposed in this thesis, and the attribute reduction algorithm based on matrix on classical information system and fuzzy information system are given, which provide a new method for knowledge acquisition of rough sets. The main research results and innovations as follows:1. To the classical information system, the relative concepts of rough set is defined by the equivalence relation, then the attribute reduction algorithms based on equivalent matrix are proposed, and the applications of attribute reduction in information table and decision table are given.2. According to the fuzzy information systems and fuzzy decision system, the relative concepts of fuzzy rough sets and rough fuzzy sets such as, upper (lower) approximation operator etc., are defined by fuzzy matrix, and the attribute reduction algorithms on them are given, moreover, the feasibility of the algorithm are verified by UCI sets.3. The method and algorithm of attribute reduction through constrcting coving from fuzzy information system are given, moreover, the feasibility of the introduced algorithm are verified by UCI database.The main idea of this thesis is characterizations the rough set theory by the matrix, which provided a new method for knowledge acquisition in fuzzy information system. The studies in this thesis enrich the relative theory of fuzzy set and rough set, which provide the foundation for the further research.