| Rough set theory, introduced by Z.Pawlak in 1982, is a powerful mathematical theory of reasoning about data after the propose of Probability theory and Fuzzy theory. It is based on the indiscernibility relation that describes indistinguishable objects. The most remarkable characteristic of this theory is it does not require any additional empirical information of data sets, it can analyze and process the non-accurate, non-integrity and incomplete data. Rough set theory is becoming a hot spot in computer science, artificial intelligence and information science.The knowledge reduction is the main topic in Rough set theory. It includes attribute reduction and attribute's value reduction. Attribute's value reduction is more simple than attribute reduction. We all pay more attention to the attribute reduction. Attribute reduction is the important problem in Rough set theory. Usually, knowledge reduction is named attribute reduction.At the first part of this paper we first discussed the limitation of the definition of accuracy measure in classic rough set theory. Then we proposed a new definition of the accuracy measure and the attribute's significance. Secondly three ways of attribute reduction in complete information system based on attribute's significance, information quantity and theirs'harmonic mean definition were proposed, through the same example we discussed theirs'difference. Thirdly we promoted the definition of accuracy measure and information quantity to incomplete information system.At the second part of this paper we discussed the definition of accuracy measure of indiscernibility in classic rough set and variable precision rough set, and a new accuracy measure way based on variable precision rough set was proposed. This new way could decrease the risk of the error in some making-decisions. Finally, based on this new calculate way, we proposed a method to construct decision tree to obtain the simplest attribute reduction of attribute sets.
|
PDF Full Text Request