Defending Digital Wallets: A Multi-Method Approach to Detect Fraudulent Card Transactions

Abstract

Credit card fraud poses a significant threat to the financial industry, causing substantial losses for both individuals and institutions. This paper presents a comprehensive investigation into the application of machine learning techniques to enhance credit card fraud detection. We explore a diverse range of algorithms, including K-Nearest Neighbors (KNN), Naive Bayes, Logistic Regression, and Random Forest, to accurately identify fraudulent transactions. To address the inherent class imbalance in fraud detection datasets, we employ the Synthetic Minority Over-sampling Technique (SMOTE). Additionally, we delve into Exploratory Data Analysis (EDA) to gain valuable insights into the data distribution and potential patterns that may aid in fraud detection. To optimize model performance, we meticulously tune hyperparameters, fine-tuning the algorithms to achieve optimal results. The result confirmed that the Random Forest model is more advantageous than the other models considered (K-Nearest Neighbor, Naïve Bayes, and Logistic Regression) because it provides high accuracy and balance in the ability to predict classes. Our empirical evaluation demonstrates the effectiveness of the proposed approach, highlighting its potential to significantly improve the accuracy and precision of credit card fraud detection systems.

IJSS, Vol. 26(1), March, 2026, pp 95-106