COMPARATIVE EXAMINATION OF DEBIT CARD FRAUD DETECTION METHODS EMPLOYING MACHINE LEARNING INTELLIGENCE APPROACHES

Abstract

For this study, we look at how to spot fake debit card transactions using supervised intelligent expert systems techniques like Extreme Gradient Boosting, Random Forest, KNN, Logistic Regression, SVM, and Decision Trees to solve classification problems. A debit card dataset with two parts, 30 variable fields, and a total of 248,807 records was used to do a thorough evaluation of data mining and machine learning algorithms. This dataset helped with the investigation of fraud detection. After finishing the basic structure, I can say with certainty that the combined technique model performs much better than a separate system that uses k-means, XGBoost, SVM, and logistic regression to find outliers. Given that the main objective is to detect fraudulent activities in a debit card dataset, its effectiveness is measured by how frequently it identifies outliers or atypical user behaviour patterns.