GreyART Network for Financial Distress Prediction Problem

Abstract

This study attempts to use the GreyART network to construct a financial distress prediction model. The inputs applied to the network are the historical data containing 18 different financial ratios of 54 healthy and 22 distressed Taiwan's listed electronic firms. In order to determine the best result the GreyART network can attain, a new performance index is developed. Simulation results show the one using 8 variables to generate only four clusters, 1 for healthy class and 3 for distressed class with corresponding classification hit rates of 94.12% and 93.55% for the training and test phases, respectively.