Banks are interested in evaluating the risk of the financial distress before giving out a loan. Many researchers proposed the use of models based on the Neural Networks in order to help the banker better make a decision. The objective of this paper is to explore a new practical way based on the Neural Networks that would help improve the capacity of the banker to predict the risk class of the companies asking for a loan. This work is motivated by the insufficiency of traditional prevision models. The sample consists of 86 Tunisian firms and 15 financial ratios are calculated, over the period from 2005 to 2007. The results are compared with those of discriminant analysis. They show that the neural networks technique is the best in term of predictability.
@article{arxiv.1311.4266,
title = {Pr\'evision du risque de cr\'edit : Une \'etude comparative entre l'Analyse Discriminante et l'Approche Neuronale},
author = {Younes Boujelbène and Sihem Khemakhem},
journal= {arXiv preprint arXiv:1311.4266},
year = {2013}
}