An Evolutionary Based Wavelet Network for Business Failure Prediction

Forecasting business failure is an important and challenging task for both academic researchers and business practitioners. A large number of methods like discriminant analysis, log it analysis, neural networks,etc., have been used in the past for the prediction of business failure. Although some of these methods lead to models with a satisfactory ability to discriminate between healthy and bankrupt firms, they suffer from some limitations, often due to the unrealistic assumption of statistical hypotheses or due to the learning problems of poor convergence in the training process. In this paper, an evolutionary based wavelet network is presented to discriminate between healthy and failing firms in order to predict business failure.Financial characteristics of a large sample of 256 Chinese firms are used to train the proposed network and to evaluate its prediction ability. The results are very encouraging, compared with those of multiple discriminant analysis, log it regression and pure neural network, and prove the usefulness of the proposed method for business failure prediction.