Application of the Cross Entropy Method to the Credit Risk Assessment in an Early Warning System

Abstract

Traditional Moonier Carlo method usually takes a long time to simulate rare event, while importance sampling techniques can effectively reduce the simulation time and improve simulation efficiency. In this paper, an importance sampling method - cross entropy is presented to deal with credit risk assessment problems for commercial banks. The failure event of repaying loans is treated as rare event due to the relatively low probability, and the failure probability of repaying loans is taken as the criterion to measure the level of credit risk. Numerical experiments have shown that the cross entropy method has a strong capability to identify the credit risk and it is a good tool for credit risk early warning system.