Abstract: Feature selection is an essential step in preprocessing and it refers to the process of selecting input variables that are most predictive for a given outcome. Reducing the input space is of major concern in areas like pattern recognition, signal processing, medical research and machine learning. Use of rough set theory for preprocessing of dataset has been very recent since other methods are inadequate at finding minimal reductions that too with uncertain data. The essence of this study, is to introduce an innovative approach by fusing rough set theory with feature correlation for reducing the input space and then applying fuzzy ART neural network for breast cancer diagnosis. The intended approach has produced satisfactory results as opposed to the conventional methods.
A. Punitha and T. Santhanam , 2007. Correlated Rough Set Based Classificatory Decomposition for Breast Cancer Diagnosis Using Fuzzy ART Neural Network. Asian Journal of Information Technology, 6: 1212-1217.