Loan Prediction System

• Tech Stack: Python, Ensemble Methods, SHAP, HTML/CSS
• Github URL: GitHub Link

• Engineered a robust loan approval framework by implementing Ensemble Methods and advanced data balancing techniques to mitigate class imbalance in banking datasets.
• Attained a test accuracy of 93.43% and a mean cross-validation accuracy of 93.23% by optimizing multi-layer neural networks and ensemble architectures.
• Accomplished seamless deployment of diagnostic insights by creating a web-based UI with HTML and CSS, allowing for real-time prediction testing and model interpretability via SHAP visualizations.