Credit Card Fraud Detection Using Machine Learning

The proliferation of e-commerce and e-payment systems has led to a surge in financial fraud cases, particularly credit card fraud. Detecting fraudulent activities is paramount in safeguarding users' financial assets and maintaining trust in online transactions. This paper presents a novel approach for credit card fraud detection utilizing machine learning (ML) techniques coupled with genetic algorithm (GA) for feature selection. Feature selection is critical in enhancing the effectiveness of fraud detection models by identifying the most relevant features associated with fraudulent transactions. The performance of the proposed approach is evaluated using a dataset sourced from European cardholders, a common benchmark dataset in the field. The Isolation Forest Algorithm has a slightly higher accuracy compared to the LOF Algorithm. Therefore, based solely on accuracy, the Isolation Forest Algorithm is considered the better algorithm for this particular dataset.