Data Mining has been an urgent need because of increasing size of current databases. Rough Set theory has become an important method for data mining due to its unique advantage in knowledge discovery. An information entropy-based uncertainty measure is presented first based on generalized rough set model in this paper, which is suitable for evaluating rules retrieved from noisy data. Second, this paper puts forward generalized minimal-and-maximal-rules-learning methods and generalized maximal-minimal-rules-conversion model because we can encounter noisy problems in most real-life problems. Third, This paper puts forward a new discretization method for the continuous attributes, which is based on the clustering and rough sets theory. The empirical result illustrates that we can simplify the single rule and reduce the number of rules through the GMM method. We can also solve the noisy problem in the practical application effectively. |