Hybrid arithmetic optimization algorithm for telecom network anomaly detection

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Telecom networks face sophisticated anomalies and intrusions.
• Existing intrusion detection systems (NID) require enhanced accuracy and efficiency.
• Network traffic analysis is crucial for maintaining network integrity.

Purpose of the Study:

• To introduce a hybrid algorithm, OBAOANM, for superior telecom network anomaly detection.
• To enhance the exploitation and exploration capabilities of the Arithmetic Optimization Algorithm (AOA).
• To improve the identification precision of network bypass anomalies and overall NID.

Main Methods:

• Hybridization of Opposition Based learning (OB) and Nelder-Mead (NM) algorithms with the Arithmetic Optimization Algorithm (AOA).
• Developing the OBAOANM algorithm to extend AOA's phases for better anomaly identification.
• Utilizing OBAOANM for network traffic analysis and intrusion detection.

Main Results:

• The OBAOANM algorithm demonstrated enhanced identification of network bypass anomalies.
• Achieved an average classification accuracy of 93.10% in comparative NID experiments.
• The proposed method proved effective in maximizing classification accuracies for NID.

Conclusions:

• OBAOANM is a robust tool for addressing network traffic problems in telecom environments.
• The hybrid approach significantly boosts the precision of telecom network intrusion detection.
• The algorithm offers a more suitable solution for complex network security challenges.