A Two-stage Intelligent Model for Personal Credit Loans Default Prediction by Using SOM and PHM

Abstract

In the era of big data, it is very important for companies to mining the value of vast data in the data warehouse. This study provides a two-stage intelligent predicted model adds the concept of customer relationship management (i.e. customer segmentation). As a result, this study applied SOM to solve the urgent issue and find 3 groups of borrowers, then we predict the lifetime of borrowers by using PHM. The results show that the borrowers can be segmented effectively into different group by SOM and the performance of integrated model is the best.