A COMPARATIVE SUPERVISED MACHINE LEARNING FRAMEWORK FOR CREDIT CARD FRAUD DETECTION ON HIGHLY IMBALANCED TRANSACTION DATA

Megha Baghsawari , Swati Choudhary , Muskan Uday , Twinkal Yadav , Deepali Chourey

Credit Card Fraud Detection, Supervised Machine Learning, Random Forest, Imbalanced Data, Financial Analytics

The rapid expansion of digital payment systems has significantly increased the prevalence and sophistication of credit card fraud, posing serious financial risks to consumers and financial institutions. Traditional rule-based fraud detection systems struggle to adapt to evolving fraud patterns and large-scale transaction data. To address these challenges, this study presents a comprehensive comparative analysis of three supervised machine learning models—Logistic Regression, Decision Tree, and Random Forest—for detecting and predicting fraudulent credit card transactions using a highly imbalanced dataset. Data preprocessing techniques, including feature scaling and undersampling, are employed to mitigate bias toward the majority class. Model performance is evaluated using accuracy, precision, recall, F1-score, specificity, and the area under the receiver operating characteristic curve (AUC). Experimental results demonstrate that the Random Forest model outperforms the other classifiers, achieving an accuracy of 96% and an AUC of 98.9%. Additionally, demographic and temporal analyses reveal that cardholders above 60 years are more vulnerable to fraud, with a higher frequency of fraudulent transactions occurring between 22:00 and 04:00 GMT. The findings highlight the effectiveness of ensemble-based learning approaches and provide practical insights for enhancing fraud detection systems in the financial sector.

Volume 10 Issue 12, December-2025

Pages : b522-b530