Cyber Attack Detection using Ensemble Deep Multi-Model Algorithm to Detect User Behavior

The malicious activities of end user’s behavior on Wireless Network has more focus on new detection of cyber threats are major studies of past one decade in research. This paper presents an Ensemble Deep Multi-Model Algorithm (EDMMA) designed to enhance cyber-attack detection capabilities. By merging the capabilities of CNNs, LSTMs, and Transformers, the hybrid framework achieves greater precision and resilience in detection tasks by capitalizing on their distinct strengths. This solution fuses CNN, LSTM, and Transformer architectures, harnessing their unique features to boost accuracy and robustness in detection. Extensive experiments conducted on large-scale user behavior datasets demonstrate that the EDMMA with 97.50% outperforms against LSTM algorithm of 90.40% in terms of accuracy, precision, recall, and F1-score, making it a promising solution for real-world cybersecurity applications.