Credit card fraud detection remains a significant challenge due to the growing complexity of fraudulent behaviour and the severe class imbalance in transaction data. This study presents a hybrid deep learning approach that combines three advanced models-an Artificial Neural Network (ANN) enhanced with Batch Normalization and Dropout, along with VGG16 and VGG19 architectures-to enhance detection accuracy and reliability. The system begins with extensive data pre-processing, including Standard Scaling for normalization, Synthetic Minority Over-sampling Technique (SMOTE) to balance class distribution, and Principal Component Analysis (PCA) for dimensionality reduction.
Kundinnen und Kunden meinen
0.0/5.0
0 Bewertungen
Verfassen Sie die erste Bewertung zu diesem Artikel