Effective Multi Class Attack Recognition through Formed Arrayized Loop with Integrated Bi-GRU

This study introduces a pioneering framework for precise multi-class attack recognition employing a Formed Arrayized Loop with Integrated Bidirectional Gated Recurrent Unit (Bi-GRU). The proposed model addresses the shortcomings of existing network intrusion detection systems (NIDS), which often struggle with lower accuracy, the inability to perform multi-class classification, increased training time,and inefficiency with large datasets. By integrating the Formed Gate Loop Model (FGLM) and Arrayized Trigger Unit Function (ATUF) within the Bi-GRU framework,the model achieves high accuracy and efficient learning. The results reveal a swift decrease in training loss from 0.2967 to 0.0732 and a corresponding increase in training accuracy from 0.9264 to 0.9828. Similarly, the validation loss decreases from 0.0905 to 0.0364, while the validation accuracy rises from 0.9627 to 0.9892 and stabilizes after the tenth epoch, signifying robust generalization capabilities. The close alignment of training and validation metrics suggests minimal overfitting and effective learning of underlying patterns. The proposed approach enhances the selectivity of updating memory, ensures important features are retained, and reduces data dimensionality, leading to faster convergence and improved prediction performance. Recommendations for future work include expanded dataset utilization, real-time implementation, hybrid model development, advanced feature engineering, adaptive learning mechanisms, and the development of a user friendly interface. This research contributes to the development of robust and effective network intrusion detection systems, essential for safeguarding modern networks against sophisticated attacks.