Feature selection via sensitivity analysis of MLP probabilistic outputs

This paper presents a new wrapper-based feature selection method for multi-layer perceptrons (MLP) neural networks. It uses a feature ranking criterion to measure the importance of a feature by computing the aggregate difference, over the feature space, of the probabilistic outputs of the MLP with and without the feature. Thus, a score of importance with respect to every feature can be provided using this criterion. The proposed criterion has inexpensive evaluation. Based on the numerical experiment on several artificial and real-world data sets, the proposed method performs at least as well, if not better, than several existing feature selection methods for MLP.