A novel research on network security situation prediction based on iteratively optimized RBF-NN.

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Network security situation (NSS) prediction is crucial for mitigating cyberattacks.
• Existing prediction methods struggle with non-stationary, non-linear data, leading to slow convergence and local optima.
• Limited generalization ability hinders the effectiveness of current NSS prediction models.

Purpose of the Study:

• To propose a novel iterative optimized Radial Basis Function Neural Network (RBF-NN) for enhanced NSS prediction.
• To address the limitations of existing methods, including slow convergence and susceptibility to local optima.
• To improve the accuracy and efficiency of NSS prediction for complex network environments.

Main Methods:

• Utilized a resource allocation network (RAN) to dynamically determine the optimal number of hidden layer neurons.
• Implemented a cross-model approach with a genetic algorithm for optimal RBF-NN weight computation.
• Incorporated a chaos search strategy to prevent the model from converging to local extreme points during optimization.

Main Results:

• Achieved a significant improvement in prediction accuracy, up to 86.6%.
• Reduced training time by up to 29.2% compared to existing techniques.
• Demonstrated efficient and effective performance for real-world NSS prediction tasks.

Conclusions:

• The proposed iterative optimized RBF-NN method offers superior performance for NSS prediction.
• The integration of RAN, genetic algorithms, and chaos search effectively overcomes limitations of previous methods.
• The method provides a computationally efficient and highly accurate solution for proactive network security.