End-to-end analysis of bank loan default risk using historical lending data to identify key risk factors, assess borrower behavior, and support data-driven credit decisions.

Uni-variate and Bi-variate analysis to understand the driving factor behind loan default

Predicting loan defaults using machine learning and hybrid feature engineering approaches.

A machine learning–based credit risk prediction system using XGBoost, deployed as an interactive Streamlit web application to classify applicants as Good or Bad credit risk.

Logistic Regression model predicting loan repayment vs default using financial attributes. Strong ROC-AUC (0.91) with business interpretability.

EDA and hypothesis testing project to identify key factors in loan default analysis

BankPrediction is a machine learning project that predicts customer churn, loan defaults, and detects anomalies using XGBoost models. It includes visual data analysis and a future prediction model for 2033.

Explainable ML system for loan default prediction integrating cybersecurity-inspired behavioral features. 99.43% ROC-AUC. Master's thesis project.

Credit risk assessment using FICO score segmentation, loan default modeling, discretization techniques, and log-likelihood evaluation for predictive analytics in financial services.

A machine learning project to predict credit risk (GOOD or BAD) for loan applicants using historical loan data from 2007–2014. This solution helps multifinance companies minimize default risk and streamline loan approvals through accurate risk classification and a modern graphical user interface (GUI).

Production-ready ML pipeline for retail banking default prediction with feature engineering, CatBoost models, and Dockerized FastAPI deployment on Google Cloud.

Predictive Modeling for Loan Default Risk using Machine Learning

A machine learning project focused on predicting the Probability of Default (PD) on loans using historical financial and demographic data. The project includes data preprocessing, feature engineering, model training, hyperparameter tuning, and interpretability using SHAP values.