in recent years, fraud technologies became more improved and easier for fraud. This is why a wide variety of machine learning methods were implemented and advanced for the recognition of the transactions of fraudulent credit cards. The fundamental issue with failing any detection technique on any fraud operation represents the results' accuracy. The present study included a discussion of the ways for improving the efficiency of fraud detection with the use of the approaches of machine learning and feature reduction with the use of the Rough Set Theory, followed by the use of the Voting approach for choosing the most suitable approach for inclusion in the systems fraud detection. It has also provided comprehensive research on the European customer database and how the classifiers interact with it by applying three classification algorithms. In addition to that, they were utilizing Python language tool to apply machine learning algorithms with the approach of Voting for choosing the optimal model of a classifier. Experimental results have shown that the DT algorithm's use is the optimal one as it has been capable of achieving 99.82% accuracy and 0.8 sec. Processing time.

in recent years, fraud technologies became more improved and easier for fraud. This is why a wide variety of machine learning methods were implemented and advanced for the recognition of the transactions of fraudulent credit cards. The fundamental issue with failing any detection technique on a...