We fitted models of Logistic Regression, Naive Bayes, Random Forest, KNN classifier, and Neural Networks to credit card consumer data to predict payment default. With Neural Networks, we attempted a novel architecture which exploited the partially sequential nature of the given data. Using both accuracy and a metric called "area ratio" to gauge prediction confidence, we compared the models to figure out which one was best.

Dataset: https://archive.ics.uci.edu/ml/machine-learning-databases/00350/