A Study on Feature Subset Selection Using Rough Set Theory

Feature subset selection is an important component of knowledge discovery and data mining systems to help reduce the data dimensionality. Rough sets theory provides a mechanism of selecting feature subsets. Most existing rough set-based feature selection algorithms suffer from intensive computation of either discernibility functions or positive regions to find attribute reduct. In order to improve the efficiency, we propose a new concept, called relative attribute dependency, with which we present a sufficient and necessary condition of the minimum conditional attribute reduct of a decision table and develop a new computation model to find the minimum reduct of condition attributes. The relative attribute dependency can be calculated by counting the distinct rows of the sub-decision table, instead of generating discernibility functions or positive regions. Thus the computation efficiency of minimum reducts is highly improved. Two novel algorithms to find optimal reducts of condition attributes based on the relative attribute dependency are proposed and implemented using Java, one brute-force algorithm and the other heuristic algorithm using attribute entropy as the heuristic function. The algorithms proposed are experimented with 10 data sets from UCI Machine Learning Repository. We conduct the comparison of data classification using C4.5 with the original their their usefulness and are analyzed for further reserach.