Predict Default of Next Credit Card Payment

Problem Statement

HCL Technologies Problem Statement 2 - credit card default

Introduction

This project intends to create a model that can predict if a customer will default the next credit card payment based on the inputs provided to the model. The accurate prediction of the next credit card default of the customer will enable financial institutions to better mitigate such risks

Purpose & Motivation

To use AI to assist financial institutions in identifying credit card customers who are not credible to allow proactive risk handling

Application Details

The model is created using Azure Machine Learning Studio. Information may be fed from other places (e.g. web services) into the model after it has been deployed and model can be automatically run. The output can then be fed to the financial institution on the customer’s classification status

Difficulties & Challenges

1. Steep learning curve for the software

2. Unable to create new account for the free trial after it expired despite creating new email addresses to do so

Go-to-Market

It can be available for use by financial institutions as a back-end application that automatically runs check on customers based on the model to constantly update the classification of their customers and flag out any potential high-risk customers (i.e. customers with higher risk of defaulting next credit card payment)

Built With

• azure
• python