Identifying Fraudulent Credit Card Transactions Using AI

Credit card fraud is a major problem that costs businesses and consumers billions of dollars each year. Machine learning techniques can be used to detect fraudulent transactions, but they often struggle with imbalanced datasets, where there are far more legitimate transactions than fraudulent ones. This research aims to develop a machine-learning model which will automatically detect credit card fraud. The performance of four machine learning algorithms for credit card fraud detection: Knearest neighbour (KNN), decision tree, random forest, and XG Boost have been evaluated. The algorithms were trained using four different methods to ensure a comprehensive evaluation. The performance of each algorithm was evaluated using accuracy, precision, F1-score, roc value and recall. The results of the study show that XGBoost and Decision Tree achieved the highest accuracy and recall among the four algorithms for the synthetic dataset. Both models were also able to achieve a high precision, which means that it was able to identify a high percentage of fraudulent transactions while minimizing the number of false positives. Overall, the findings suggest that XG Boost and Decision Tree are promising machine learning algorithms for credit card fraud detection. However, it is important to validate their performance on diverse datasets to assess their generalizability and robustness.