A global discretization method based on rough sets

Since rough sets theory can unveil the dependency of data and implement data reduction, it has been increasingly researched in more and more fields. In rough sets theory and other induction learning systems, discretization is an important algorithm and can be viewed as a process of information generalization (or abstraction) and data reduction. In this paper, a global discretization algorithm is proposed based on rough sets. It modifies the criterion of selecting the best cut points, and introduces inconsistency checking to preserve the fidelity of the original data, which change the MDLP method into a global one. Thus the reduction of cut points can be performed while keeping the consistency level. The proposed algorithm is tested by using several data sets with ID3 and ROSETTA. Experiments results show that this method performs better than MDLP, and is also superior to those which process continuous data directly without discretization.