Knowledge discovery in database(KDD), a new generation of tools and techniques for automatic and intelligent database analysis, is an active area with the promise for a high payoff in many business and scientific application. It is defined as the nontrivial process of identifying valid, novel, potentially useful and ultimately understandable patterns in data. Data mining is the core step of the knowledge discovery. The theory of Rough Sets, presented in 1982 by polish mathematician-Z. Pawlak, is a powerful mathematical tool for analyzing uncertain, unclear data and is widely applied to knowledge discovery or data mining.The following two aspects are focused on in this paper. One is the mathematical model of the data mining based on the Rough Sets, and the other is the attribute reduction algorithm based on the Rough Sets.Traditional mathematical model of data mining based on the Rough Sets can not be applied to deal with the database directly. For the sake of this, an improved mathematical model of data mining based on the Rough Sets is presented in this paper. Attribute reduction algorithm is the key for the model and the focus of the Rough Sets. For the time being, many reduction algorithm are presented, however, most of them are inefficient or impractical. On the basis of the research on Rough Sets theory and known reduction algorithms, an improved attribute reduction algorithm is presented in this paper. This heuristic, improved attribute reduction algorithm, based on the Hu's algorithm and Jelonek's algorithm, can guarantee a reduction compared with the Hu's algorithm, and on the other hand, has better efficiency as far as time is concerned compared with the Jelonek's algorithm. |