Bank Loan Prediction by using Logistic Regression

Abstract

In the modern era of science and technology, the banking sector is also growing very fast and a lot of people are applying for bank loans. But the bank has limited assets and can be granted to limited number of people only, so the bank must ensure that the person whom the loan is granted is eligible enough to pay it back. So, in this study, we will try to reduce the risk factor involved in finding the eligibility of the client so that the bank can reduce its effort in selecting the client. This model is based on Big Data mining of previous records of people to whom the loan was granted. The model is trained based on these experiences of the people in the record and various machine learning algorithms are used and a suitable algorithm is selected for better accuracy. The main objective of this study is to predict whether a particular person is eligible for loan or not. This research work contains four sections: (i) Data Collection and Pre-processing (ii) Comparison of machine learning models based on accuracy (iii) Training of system on the most promising machine learning model (iv) Testing the model. In this study we analyze and predict the loan data by using various algorithms: Linear Discriminant Analysis, Random Forest, Logistic regression, Gaussian NB, Decision Tree and SVC.