Network Security Situation Prediction Based on Optimized Clock-Cycle Recurrent Neural Network for Sensor-Enabled...

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Network security situations exhibit complex temporal dynamics and nonlinearity.
• Accurate prediction and real-time monitoring are crucial for effective network security.
• Existing models may struggle to capture both short-term and long-term dependencies efficiently.

Purpose of the Study:

• To propose an optimized Clockwork Recurrent Neural Network (CW-RNN) for network security prediction.
• To enhance the model's ability to capture temporal features and nonlinear dynamics.
• To improve prediction accuracy and real-time performance in network security monitoring.

Main Methods:

• Implemented a Clockwork Recurrent Neural Network (CW-RNN) architecture.
• Utilized a clock-cycle mechanism within hidden units to process information at different frequencies.
• Employed the Grey Wolf Optimization (GWO) algorithm for hyperparameter tuning.
• Evaluated the model's performance on network security situation data.

Main Results:

• The optimized CW-RNN demonstrated superior performance in extracting temporal and nonlinear features.
• The model achieved improved prediction accuracy compared to other network models.
• The approach exhibited low time complexity and excellent real-time performance.
• Effectively captured both short-term and long-term temporal dependencies in network data.

Conclusions:

• The proposed CW-RNN with GWO optimization offers an effective solution for network security prediction.
• The clock-cycle mechanism enhances the model's capability to learn complex temporal patterns.
• The approach is suitable for real-time monitoring of large-scale network traffic, particularly in sensor networks.