IEEE TRANSACTIONS ON SYSTEM , MAN , AND CYBERNETICS { PART B : CYBERNETICS 1 The ANNIGMA - Wrapper Approach to Fast FeatureSelection for Neural

|This paper presents a novel feature selection approach for backprop neural networks. Previously, an feature selection technique known as the wrapper model was shown eeective for decision trees induction. However, it is prohibitively expensive when applied to real-world neural net training characterized by large volumes of data and many feature choices. Our approach incorporates a weight analysis based heuristic called ANNIGMA to direct the search in the wrapper model and allows eeective feature selection feasible for neural net applications. Experimental results on standard data sets show that this approach can eeciently reduce the number of features while maintaining or even improving the accuracy. We also report two sucessful applications of our approach in the helicopter maintenance applications.