A rough sets based approach to feature selection

Feature selection techniques aim at reducing the number of unnecessary features in classification rules. The features are measured by their necessity in heuristic feature selection techniques. Rough set theory has been used to define the necessity of features in literature. We propose a new rough set based feature selection approach called Parameterized Average Support Heuristic (PASH). The PASH considers the overall quality of the potential set of rules. It selects features causing high average support of rules over all decision classes. In addition, the PASH arms with parameters that are used to adjust the level of approximation.