CORPORATE DISTRESS DIAGNOSIS - COMPARISONS USING LINEAR DISCRIMINANT-ANALYSIS AND NEURAL NETWORKS (THE ITALIAN EXPERIENCE)

This study analyzes the comparison between traditional statistical met hodologies for distress classification and prediction, i.e., linear di scriminant (LDA) or logit analyses, with an artificial intelligence al gorithm known as neural networks (NN). Analyzing well over 1,000 healt hy, vulnerable and unsound industrial Italian firms from 1982-1992, th is study was carried out at the Centrale dei Bilanci in Turin, Italy a nd is now being tested in actual diagnostic situations. The results ar e part of a larger effort involving separate models for industrial, re tailing/trading and construction firms. The results indicate a balance d degree of accuracy and other beneficial characteristics between LDA and NN. We are particularly careful to point out the problems of the ' black-box' NN systems, including illogical weightings of the indicator s and overfitting in the training stage both of which negatively impac ts predictive accuracy. Both types of diagnostic techniques displayed acceptable, over 90%, classification and holdout sample accuracy and t he study concludes that there certainly should be further studies and tests using the two techniques and suggests a combined approach for pr edictive reinforcement.