Optimising Security: A Hybrid CNN-BiLSTM Model for Credit Card Cyber Fraud Detection

The banking industry has long recognised the importance of developing efficient credit card cyber fraud detection systems. Despite these efforts, businesses face rising credit card cyber fraud due to technological development. The growing frequency of cybersecurity data breaches has made credit card cyber fraud detection systems less efficient at detecting advanced fraud. Although Machine Learning(ML) algorithms have been employed to detect credit card cyber fraud, no cyber fraud detection system has yet to achieve a high level of efficiency. Using DL techniques for credit card fraud detection has yielded better performance than traditional algorithms. It tackled the problem of detecting unexpected and sophisticated cyber fraud patterns. This paper proposes a novel Hybrid CNN-BiLSTM model to effectively address credit card cyber fraud. The novel hybrid model uses DL techniques like convolutional neural network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM). We conducted a thorough comparison of the accuracy, precision, recall, Fl score, and ROC-A UC score. The experiments demonstrate that the innovative CNN-BiLSTM model surpasses the performance of each individual model. The integration of CNN and BiLSTM techniques represents substantial progress in the domain of cyber fraud detection, thereby boosting the security offinancial transactions.