Monitoring and crediting process control with use of logical and probabilistic risk model

We suggest the new technology of monitoring and crediting process control for a commercial bank. A logical and probabilistic (LP) risk model for the credit risk is described and its advantages are demonstrated. We prove the impossibility to create identical learning and test samples with the same frequencies of events for the LP risk model. The offered technology includes: determination of the minimal volume of the statistical data for the learning sample, exception of the part of outdated and incorrectly recognised credits in the statistical data, introduction of the signal set of the finalised credits, periodical re‐learning and substitution of the LP risk model, and crediting process control.