Performance Review of Popular Classification Models

In this paper, we examine and compare the performance of three popular models used for classi▯cation: logistic regression, neural networks and XGBoost. To quantify performance using a variety of metrics, we fit and test the models on credit loan data provided to the public by Lending Club. We ▯nd that the XGBoost algorithm is the clear winner as it performs best even with minimal hyperparameter tuning.