Enhancing Credit Card Fraud Detection: A Comparative Analysis of Artificial Neural Networks, k-NN, and Naive Bayes Classifiers

This paper conducts a comparative study of three machine learning algorithms—artificial neural networks (ANNs), k-Nearest Neighbors (k-NNs), and Naive Bayes (NB)—for irregular payment card detection in currency a to pay for the growing incidence of card fraud In the digital age. Each algorithm presents unique strengths and challenges. ANNs excel in capturing complex nonlinear relationships, while k-NN shows efficiency in well-defined data sets, and NB for elegant probabilistic modeling The study explores hybridization techniques and clustering techniques to benefit from complementary nature of these algorithms. It also highlights the importance of preprocessing methods such as feature engineering and imbalanced class control to prepare raw data for advanced learning in addition if dynamic learning methods are deemed to provide learning material management is well developed.