This repository contains the master code file and data-set used for the credit risk modeling session in Intelligent Machines and shared in https://www.youtube.com/watch?v=oglRV595584
Tools in Predictive Analytics and Machine Learning are gaining widespread acceptance in many different applications, and times in areas with already established existing methodologies. Such an area is Credit Risk Modeling, which is entrenched in Financial Rations and Calculations. This code, and the video tutorial, aims to present the task of CR Modeling using purely Machine Learning Tools. We would like to argue that both these analyses are not substitutes to existing mechanisms, but actually can supplement them to make better credit default predictions.
- Exploratory analysis of the data and data viz
- Regression Models of Predictions
- Understanding drivers to default rate
- Measuring Model Performance for different Regression Models
- Recursive Partitioning methods and their Performance
- Analyzing models based on Acceptance Rate of credit applications and Bad Rate
- Random Forest and Performance
- Neural Network and Performance
We are welcome to hear from you if you have any ideas on CR Modeling. One very interesting application will be to use these mdoels side by side with financial metrics of a bank, and see if these can support in improving the CR prediction performance.
A report highlighting the insights!
Enjoy!