Predictive Modeling for Personal Loan Acceptance in Banking

Hello everyone,

From past few days I was working on a project which is based on the data mining for the personal loan acceptance in the field of banking sector. The data which I have collected is of the Thera Bank and the data shows the major attributes such as age, income category, mortgage, securities account, CD account, and credit card history.

In this analysis, I performed the Logistic Regression and CART (Classification And Regression Trees). The aim of this research is to evaluate the predictive power of decision tree classification models and logistic regression models in relation to liability customers’ likelihood of applying for personal loans from “Thera” Bank. The dataset comprises demographic information, banking relationships, and responses from 5000 customers to a previous personal loan campaign. The models are tested and trained using 80/20 and 50/50 data splits. Decision tree and logistic regression categorisation models are applied and compared. Particularly, specificity and sensitivity are used to illustrate the performance metrics used to evaluate the efficacy of classification models in the context of binary classification tasks. For the purpose of metric interpretation, the AUC-ROC curve displays the TPR for sensitivity and the TNR for specificity.

This report explores the domain of data-driven insights in the banking sector, with a particular emphasis on the effective identification of prospective customers who are inclined to accept offers of personal loans. The attached paper report is below: