ENHANCED CREDIT CARD FRAUD DETECTION USING DEEP LEARNING ENSEMBLE AND DATA RESAMPLING

Abstract

Credit cards play an essential role in today’s digital economy, and their usage has recently grown 
tremendously, accompanied by a corresponding increase in credit card fraud. Machine learning (ML) algorithms 
have been utilized for credit card fraud detection. However, the dynamic shopping patterns of credit card holders 
and the class imbalance problem have made it difficult for ML classifiers to achieve optimal performance. In order 
to solve this problem, this paper proposes a robust deep-learning approach that consists of long short-term memory 
(LSTM) and gated recurrent unit (GRU) neural networks as base learners in a stacking ensemble framework, with a 
multilayer perceptron (MLP) as the meta-learner. Meanwhile, the hybrid synthetic minority oversampling technique 
and edited nearest neighbor (SMOTE-ENN) method is employed to balance the class distribution in the dataset. The 
experimental results showed that combining the proposed deep learning ensemble with the SMOTE-ENN method 
achieved a sensitivity and specificity of 1.000 and 0.997, respectively, which is superior to other widely used ML 
classifiers and methods in the literature. Next we introduce advanced ensemble models, including Stacking and 
Voting Classifiers, evaluating them on both original and SMOTE-ENN datasets. Additionally, a Flask framework 
with SQLite integration enables user signup, signin, and testing for enhanced project functionality and user 
interaction.