This is a dataset from TVS finance. Goal is to predict the probability of loan default based on several factors Various details like loan amount, no of payments, default history etc is given of 120k customers. Based on these the future probability of defaulting a loan is predicted.
Logistic regression, KNN, Decision tree & SVM is used The data set has very less number of loan defaulters compared to non-defaulters. SMOTE method is used to balance the defaulters and non-defaulters to improve the model performance.