An Intelligent Agent-Based Detection System for DDoS Attacks Using Automatic Feature Extraction and Selection.

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Network Security

Background:

• Internet services face significant threats from Distributed Denial of Service (DDoS) attacks, advanced persistent threats, and malware, compromising availability and security.
• Existing machine learning-based DDoS detection techniques require improvement in accuracy and processing speed.

Purpose of the Study:

• To propose an intelligent agent system for enhanced detection of DDoS attacks.
• To improve the accuracy and processing speed of DDoS attack detection methods.

Main Methods:

• Development of an agent-based mechanism combining machine learning and sequential feature selection.
• Utilization of automatic feature extraction and selection for dynamic agent reconstruction.
• Experimentation using the CICDDoS2019 dataset, a custom-generated dataset.

Main Results:

• The proposed system achieved a 99.7% improvement over state-of-the-art machine learning-based DDoS detection techniques.
• The system demonstrated advanced detection accuracy and faster processing compared to current standards.
• Effective selection of optimal features and dynamic reconstruction of the DDoS detector agent were achieved.

Conclusions:

• The intelligent agent system offers a highly accurate and efficient solution for DDoS attack detection.
• Automatic feature extraction and selection are crucial for optimizing agent performance in real-time.
• The proposed method sets a new benchmark for DDoS detection accuracy and speed.