A comparison of neural network and Bayes recognition approaches to the diagnostic of multiple sclerosis

Abstract

This article describes the application of a Multi-Layer Perceptron (MLP) to the problem of diagnosing Multiple Sclerosis (MS). The classification information is obtained by a Trigeminal Evoked Potential (TEP) test. The performance of the MLP is compared with that of the human experts and the Bayes classifier. The efficiency of the neural network and the classical classifiers in conjunction with 4 types of features - the Fourier transform (FT), the peak position, the ARX model coefficient and the temporal wave form - are examined. Although a large clinical data base would be necessary, before this approach can be fully validated, the initial results are very promising. The MLP was found to be less susceptible to the number of features used. The ability of the MLP classifier to generalize is far better than that of the Bayes classifier.