Optimized hybrid CNN-LSTM-GRU model for cyber attack detection using Adaptive Equilibrium Optimization

Modern network security has a great difficulty in cyber attack detection. In this context, this work presents an CNN-LSTM-GRU Model for cyber attack detection. AEO was used to maximize feature selection for the model, therefore lowering unnecessary data while maintaining important attack patterns. With 98% accuracy, the suggested model beats conventional GRU, LSTM, and RNN architectures according to experimental data. Computationally efficient for the proposed model achieves equivalent accuracy to the Transformer model with less number of FLOPs and parameters.