Research Of Reduction Algorithms Based On Rough Set Theory

With the development of computer technology, people are confronted with more and more data in information age. How to find the internal information in data is a hot point with which people are concerned. As traditional DBMS can't afford the requirement which people want to extract hidden information, conceptions of Knowledge Discovery in Databases and Data Ming are expounded. Knowledge Discovery in Databases is the nontrivial process of identifying valid, novel, potentially useful and ultimately understandable patterns in databases. Data Mining is the core step during the course of Knowledge Discovery in Database. At present, it is a quite active research field.Rough Set theory, presented in1982 by Polish mathematician Pawlak, is a powerful mathematical tool for analyzing fuzzy, uncertain knowledge. As a classified method of Data Mining, Rough Set theory can effectively deal with the analysis and reduction of incomplete, uncertain knowledge, it also can find the implied knowledge and discover the latent rule. In the research of rough set theory, data reduction is one of the most important research directions of Rough Set theory. The research of data reduction based on Rough Set has greatly theoretical and realistic meaning.The thesis introduces the basic conceptions and the important laws based on the algebra view and the information view of Rough Set theory. The algebra view and the information view of Rough Set theory are analyzed and compared with each other systematically. The typical attribute reduction algorithms and their examples are discussed in the thesis .Two new attribute reduction algorithms based on the algebra view, namely EDMBARK algorithm and PRDCBARK algorithm, are developed in the thesis. Besides, two new algorithms are analyzed through theory and experiment. Explicit examples are provided to explain algorithms. Two new algorithms are beneficial to obtain the more efficient attribute reduction algorithms. At last, a improved value reduction algorithm is also discussed in the thesis.