Attribute Reduction in Information Systems via Oriented Association Coefficient

Abstract

Rough set data analysis has recently become a routine method in categorical data analysis. One of the important problems in rough set theory is attribute reduction. In this paper, the statistic named oriented association coefficient among attributes of information systems is introduced, which measures the non-linear relationships between qualitative variables. Based on this statistic, we present a new discriminant theorem of attribute reduction in information systems. Furthermore, experiments show that our approach is feasible and efficient.