A Novel Approach for Effective Detection and Prediction of Sophisticated Cyber Attacks Using the Stacked Attention GRU and BiLSTM
L Shammi, C. Emilin Shyni · 2024
The dynamic and sophisticated cyber threats of today's quickly expanding cybersecurity landscape are surpassing the effectiveness of traditional security solutions. Organizations must take a proactive stance in danger identification and mitigation if they are to successfully tackle these new difficulties. Threat intelligence (TI), or the real-time sharing of cyber threat data, is crucial for both proactive defense and quick reaction to cyberattacks. This study describes a comprehensive approach that includes feature selection, data preparation, and model training for threat intelligence-based cyber security. Normalization and standardization are two preprocessing methods that maximize data representation for efficient analysis. Principal component analysis is one of the feature selection techniques used to find the most pertinent variables for model training. A Stacked BiLSTM-A-GRU model is trained with careful respect to feature selection, guaranteeing optimal performance. A comparative examination shows that the suggested method outperforms the most advanced algorithms, with a remarkable accuracy rate of 98.71% This study emphasizes how crucial it is to use threat intelligence to have a strong cybersecurity defense against ever-changing threats.