Continuous Value Attribute Decision Table Analysis Method Based on Fuzzy Set and Rough Set Theory

In rough set theory, decision table is a kind of especial and important knowledge system which has been applied in the decision support and data mining fields widely. But rough set method can only deal with the dispersed value attribute decision table advantageously. Therefore, rough set method is limited to the analysis of discrete value attribute decision table. A key problem of the analysis of continuous value attribute decision table is to partition the continuous quantitative attribute. In this paper, combining the fuzzy set and rough set theory, a reducing method of decision table oriented to continuous value attribute is presented. In the method, continuous value attribute decision tables are dispersed via the modified FCM algorithm based on genetic optimization, so fuzzy decision tables are built, and then decision tables can be reduced easily based on rough set method. The example shows that the method is feasible and effective.