Consensus Based Bank Loan Prediction Model Using Aggregated Decision Making and Cross Fold Validation Techniques

Abstract

Considering the growth of the credit businesses, machine learning models for granting loan permissions with the minimum amount of risk are becoming increasingly popular among banking sectors. Machine Learning based models has proven to be useful in resolving a variety of banking risk prediction issues. ML Predictions are sometimes unfair and biased because they are heavily dependent on randomly selected training data sample for every prediction made. However, this problem can be address by utilizing a cross-validation strategy. Prediction can be improved by combining decisions from different machine learning algorithms (ensemble decision making). The proposed consensus-based prediction model is evaluated using standard performance metrics, and the proposed model achieved an accuracy of 83 percent.