Abstract (EN)

The design and topology of a Neural network is still an important and difficult task. To solve this problem, new approaches are proposed, and specially a combination of Induction rules with a statistical estimation of the coefficients. This research aims to compare an algorithm of this NSL approach with traditional methods (Regression and classical BP neural net). An application on the price-quality relationship for the English automobile market drive to the conclusion that the claimed superiority of the approach is not validated as, compared to BP Neural net and even linear regression, the performance of the GMDH method is inferior.