Loan_Predictions_Algorithm_Comparison

This project takes a dataset of people's loan applications and compares various algorithms and data balancing techniques to see which one has the best performance in terms of accuracy, precision, recall and F1 score. The algorithms implemented are XGBoost, Logisitic Regression and Random Forest Classifier. Each algorithm uses two different data balancing techniques (K-Fold Cross Validation and Stratified K-Fold Cross Validation). For the XGBoost implementation, we implement it in two different ways. The first way it is implemented is by method used in this research paper from UCLA. In the UCLA paper, they used different data balancing techniques so in order to improve on their implementation of XGBoost, we use K-Fold Cross Validation and Stratified K-Fold Cross Validation as data balancing techniques to see if these data balancing techiques will help improve performance. The second implementation of XGBoost uses a more standard approach with less parameters. Logistic Regression and Random Forest Classifier algorithms are also implemented to compare against.