The Research Of Attribute Reduction With Algorithm Based On S-rough Set Theory

Rough set theory is a mathematic tool to deal with fuzzy and uncertian knowledge, which is put forward by Pland mathematician Z.Pawlak in 1982. It makes use of the equivalence relations to measure the indetermination degree of knowledge and it doesn't need any knowledge beside the data which needs to be processed. Therefore the error caused by subjective appraisal can be avoided. At present, the rough set theory have been much successful applications in data analysis, pattern recognition, information science, management science, finance, medicine, chemistry ect.Application of Z. Pawlak's rough set theory to knowledge discovery in information system are some limitations: Knowledge Discovery is a closed information system, and it deals with the object which property are known (static), and the conclusion derived from information system only used to the object. While in many real applications, it often encountered with dynamic set, that is open systems, for dynamic information systems, Professor Shi Kaiquan proposed S-rough sets (Singular Rough Sets), it provides the theory basis for solving dynamic system identification, dynamic system decision, dynamic system inference and so on.Attribute Reduction of Rough Set Theory is one of the core issues, through the attribute reduction, unnecessary property are deleted from decision-making table, under the premise of without loss of basic information. These are just people's expectation. Knowledge discovery methods are mostly "static", that is mining data from potential rules which made by the relevant judge. In fact, there appears two issues in the process of recognizing a certain object. On one hand, it always lead to people's recognization partial and one-sided for the limitations of people's ability to recognize and the used tools and the demand of the recognization's result is not high for some consideration; On the other hand, though some attribute which have not impact on the rule may be reduced because of the unceasing development of the object itself, the attribute don't mean unimportant. Therefore, after some time, if a new data added, the initial result on the object may not be able to be recognized the characteristics correctly. By the general knowledge discovery methods, It is not only dissipate efforts and reduce the efficiency, but also not feasible for all data knowledge to be refound. By the view of Skowron's distinguish matrix in this paper, a distinguish matrix of S-rough set is proposed, and a attribute reduction algorithm for finding effective reduction is given.In this paper, the following work has been done:(1)The basic knowledge of the rough set theory and some core concepts are introduced in the first section. That has laid the foundation for later attribute reduction algorithm. (2) The basic knowledge of S-rough set theory is discussed. One direction attribute transfer set and two direction attribute transfer set are defined. One direction dynamic information system model and two direction dynamic information system model are studied.(3)By the view of Skowron, a distinguish matrix model of S-rough set is given, distinguish between matrix model on one direction migration and two direction migration for the dynamics of property information systems are designed respectively.(4)Distinguish matrix model on S-Rough Set and its nature are studied. Attribute reduction with algorithm based on S-rough set theory is proposed by using these nature. Decision information system's core attribute and all the attribute reduction set can be obtained by the algorithm.(5) Based on S-rough set theory, attribute reduction algorithm is applied to the synthesis of chemicals and medical treatment which prove the effectiveness of the algorithm.This article innovation is as follows:(1) Distinguish matrix model of S-rough set is given, distinguish between matrix model on one direction migration and two direction migration for the dynamics of property information systems is designed respectively.(2)Attribute reduction with algorithm based on S-rough set theory is proposed, decision information system's core attribute and all the attribute reduction set can be obtained by the algorithm.Many other problems in the decision analysis for S-rough set theory need to discuss. The according work will be done in the future.