Machine learning-based variable selection for clustered credit risk modeling

HIGHLIGHTS

• What: The authors propose a machine learning-based variable selection method based on theoretical and regulatory considerations. Through a comparative analysis using two real-world credit default datasets the authors show that the proposed approach to clustered modeling leads to the highest prediction accuracy among various clustering models. In this approach, borrowers are clustered based on their similarities through cluster analysis and for each resulting cluster a separate predictive model is developed. Because the borrowers in each cluster have similar risk characteristics, the models can be individually developed and fitted to each cluster, resulting in higher predictive . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.