Swindle: Predicting the Probability of Loan Defaults using CatBoost Algorithm

Predicting the probability of loan defaults is essential for financial institutes and banks, as a major part of their income is dependent on the interest & EMIs generated on the repayment of the loans issued by them to their customers. Most of the loans issued have a high interest rate associated with them due to lack of securities and uncertainty possessed by the customers. Hence, having a model that could predict loan defaulters would be very beneficial for the financial institutes and banks for notifying them to approve a customer's loan or not. Such a model will evaluate their customer's data based on certain parameters and generate an accurate result based on that evaluation. Swindle implements CatBoost algorithm is used for predicting loan defaults along with a document verification module using Tesseract and Camelot and also a personalized loan module, thereby mitigating the risk of the financial institutes in issuing loans to defaulters and unauthorized customers.