Defeating the Credit Card Scams Through Machine Learning Algorithms

Credit card fraud is a significant problem that is not going to go away. It is a growing problem and surged during the Covid-19 pandemic since more transactions are done without cash in hand now. Credit card frauds are complicated to distinguish as the characteristics of legitimate and fraudulent transactions are very similar. The performance evaluation of various Machine Learning (ML)-based credit card fraud recognition schemes are significantly pretentious due to data processing, including collecting variables and corresponding ML mechanism being used. One possible way to counter this problem is to apply ML algorithms such as Support Vector Machine (SVM), K nearest neighbor (KNN), Naive Bayes, and logistic regression. This research work aims to compare the ML as mentioned earlier models and its impact on credit card scam detection, especially in situations with imbalanced datasets. Moreover, we have proposed state of the art data balancing algorithm to solve data unbalancing problems in such situations. Our experiments show that the logistic regression has an accuracy of 99.91%, and naive bays have an accuracy of 97.65%. K nearest neighbor has an accuracy is 99.92%, support vector machine has an accuracy of 99.95%. The precision and accuracy comparison of our proposed approach shows that our model is state of the art.