تقدير مخاطر الائتمان المصرفي الجزائري باستعمال نماذج الانحدار اللوجيستي (LR)‎ والشبكات العصبية الاصطناعية (MLP)‎

Abstract EN

This study aims to estimate the bank loan risk in Algerian banks and make the right lending decision using logistic regression and MLP artificial neural network models.

To achieve this, we formed a database of financial ratios and qualitative variables for 100 of borrowing institutions From the CPA bank, this sample divided into 50 good Borrower institutions and 50 other bad borrower institutions.

The results of the study showed that the artificial neural networks MLP had shown an accuracy in classification at 100%, outperforming the logistic regression model whose correct classification rate of institutions was estimated at: 96%, which would help Algerian banks to predict credit risks and make wise and speed lending's decision more than the classic models, but these modern approaches require robust technologies devices and quantitative and statistical methods