Business Situation

Credit risk assessment models are used in the finance industry to create a balance between the ratio of non-performing loans (NPL) and the approval ratio. Our client wanted to renew all credit risk assessment models to decrease their NPL Rates and increase their profitability. With this purpose, they wanted to include behavioral components into their models and integrate various scoring modules as a unique score.

Our Approach & Solution

Creating an effective credit risk model requires building enriched features by combining statistical methods with domain expertise and get the best out of it. In the light of this knowledge, we have first developed a new set of variables and then used modelling techniques to increase the model performance of the existing models. With the implementation of our solution, our client will be able to differentiate more effectively between non-performing and performing customers and manage provisions and capital calculations effectively.

Technologies

We implemented our model pipeline via SAS Enterprise Miner and SAS Enterprise Guide.