Comparative analysis of artificial neural network models: Application in bankruptcy prediction

This study compares the predictive performance of three neural network meth ods, namely the learning vector quantization, the radial basis function, an d the feedforward network that uses the conjugate gradient optimization alg orithm, with the performance of the logistic regression and the backpropaga tion algorithm. All these methods are applied to a dataset of 139 matched-p airs of bankrupt and non-bankrupt US firms for the period 1983-1994. The re sults of this study indicate that the contemporary neural network methods a pplied in the present study provide superior results to those obtained from the logistic regression method and the backpropagation algorithm.