Credit Card Fraud and Bankruptcy Detection Using Machine Learning

Abstract

The study details the implementation of two separate machine learning projects referred to as "Credit Card Fraud Detection" and "Bankruptcy Detection," each of which makes use of four unique machine learning models. The research was carried out with the assistance of the K-Nearest Neighbor, Decision Tree, Support Vector Machine, and Logistic Regression models. When compiling the supervised datasets that we used, we used information from two reliable sources. We have under-sampled the data to build a more representative sample, and we have employed the feature selection approach to increase the accuracy with which these models anticipate our main goal. In the end, we assessed the results that each model produced against one another and chose the one that had the greatest overall performance. In addition to this, we have provided suggestions on how future models and data collecting may be improved.