Applied Industrial Machine Learning Towards Good Banking Credit Score Abiding The CIBIL Guidelines

Abstract

The applications of Machine Learning (ML), and it’s great capabilities to reduce strenuous functions and provide more accurate and robust results, has lured banking companies to adopt this approach, in order to keep ahead of the market competitions and gain an upper hand technologically thereby reducing the credit risks. In this rapidly advancing environment, it has become a necessity to understand the Customer’s Credit Worthiness for financial organizations, as this instills confidence in the lenders. In the past, many models and tools were developed to coup up with the risk factor concerning the banks, these tools and models were mainly statistical tools with core algorithmic concepts. Now in this ever-expanding domain of the Financial technology domain, the automation of these tasks can be achieved using classification algorithms that aid in labeling the customers based on the data fed. The paper aims to establish reliable  models and evaluate their performance reports, of the multiple machine learning models, for selection purpose, the K Fold Cross- validation technique for model selection is used, and furthermore the paper tries tries to enhance the model’s parameters for the utmost results,Anaconda Jupyternotebook is the platform used for coding and evaluating the models.To train and test the model’s accuracy,a data-set of credit card related features  are  used.  The target variable to  be  classified  is  termed  as,  ‘probability of default’, bearing in mind the utilization of the CIBIL guide framework for feature evaluation and understanding.