A Comparative Analysis for Loan Approval Prediction using Machine Learning

The financial sector of any nation is the key factor in elevating the economy of nation. The biggest issue that finances departments are facing is to approve loans for valid candidates only. Our proposed model for predicting the loan approval for valid candidates uses three different supervised learning algorithms for classification of loan approved or not, on the basis of candidate attribute values. In proposed study ensemble technique has been used to find out best classifier algorithm to train our model. In this paper we are comparing KNN, Decision Tree, Logistic Regression, SVC, and Naïve Bayes classifier algorithms based on their accuracy for provided data set to find out the highly accurate and efficient algorithm for designing a model for approving loan to valid candidate which will also help in minimizing the financial frauds.

I. Introduction

In the era of automation, different models are been designed by researchers in every domain to reduce the human efforts and increase the productivity of human beings. So far ‘n’ number of studies have been proposed on automated tasks in different disciplines of computing science. This proposed study is mainly focused on finding the highly accurate classifier algorithm to train our model for loan approval prediction. Many bank are facing issues of fraud loan application approvals. As ‘n’ number of applicants are applying for financial loan daily, so it becomes very tedious job for bank employees or finance employees to verify every application manually. So to reduce their work and to make the loan approval process more efficient and fast to handle the increased loan application, automated systems are required in finance departments. These automated systems could be designed using machine learning algorithms to train the model with labelled dataset.