Detection of Credit Card Fraud with Machine Learning Methods

The rapid growth of the world economy and the digitalization of the financial sector have brought new security challenges. The widespread use of credit cards has become a particularly convenient target for fraud. When people shop on insecure networks or websites, their card details can be compromised by fraudsters. When shopping in stores, card information can be copied with the devices placed in POS devices. In this case, fraudsters can use card information to cause financial losses. Banks attach importance to the detection of these fraudulent transactions in order to minimize the impact of these frauds on their customers. This study evaluates different machine learning (ML) algorithms for credit card fraud detection and aims to create the best performing model. The research was conducted on a dataset of credit card transactions in Europe. In the present study, ML methods such as Random Forest, Support Vector Machines, Logistic Regression, K-Nearest Neighbor, Naive Bayes and Decision Trees are evaluated. The SMOTE technique is applied to balance the minority class and the performance of the classifiers is analyzed according to metrics such as accuracy, precision, sensitivity and F1 score. The results revealed that the Random Forest algorithm performed the best with an accuracy of 99.98%. This study proves the effectiveness of data science and artificial intelligence techniques in combating fraud in the financial sector. The results can be considered as a reliable source for improving the security of financial transactions and minimizing fraudulent activities.