تطور تقدير خطر القرض في ظل نماذج الذكاء الاصطناعي

الملخص EN

Credit risk assessment is one of the most complicated issues in financial risks for banks.

This is due to both complexity and difficulty in measuring risks and building models.

Therefore, the availability of an optimal model that can measure risks related to granted credits, within the ex-ante and ex-post process of decision-making, is so crucial.

Based on purely statistical tools, conventional models have failed to a great extent to avoid the non-performing loans that had been increasing in the past.

Artificial intelligence models, though their applications are multiple beyond the financial arena, contributed effectively in this domain, in order to alleviate the exposure to credit risks.

On the basis of this assumption, the present paper tries to identify the artificial intelligence models' effectiveness used to assess credit risk with focus on the artificial neural networks (ANN) and expert systems (ES).

These models can be more important compared to conventional ones but they are still exposed to criticisms because of the shortcomings that they may have.