AzureML Based Analysis and Prediction Loan Borrowers Creditworthy

Abstract

In the era of big data, it would be beneficial for lenders to use modern technology such as machine learning algorithms to analyze and predict customers' creditworthy. In this research, our aim is to analyze LendingClub dataset to make it well understood dataset features. Then, we upload our clean dataset to Microsoft Azure machine learning (AzureML) platform to use for building our model. Which aims to predict whether the customers are going to pay back their loans or not. This model predicts the loan status going to be default or fully paid. Moreover, the LendingClub dataset we used in this work is gathered from 2007 to 2018 used accept loans. We used AzureML platform with Two Jungle algorithm and the Two Decision tree. Thereafter, we assess their performance (algorithms) in terms of Accuracy, Precision, Recall, F1 and AUC. Finally, we compare our work with other researchers and our work shows a good result compared to others.