Network anomaly detection system with optimized DS evidence theory.

Area of Science:

• Computer Science
• Network Security
• Data Science

Background:

• Network anomaly detection is crucial due to rapid network development.
• Existing methods using Dempster-Shafer (DS) evidence theory have limitations in performance and handling network complexity.
• The varied and complex nature of modern networks necessitates advanced detection techniques.

Purpose of the Study:

• To develop a novel network anomaly detection system with enhanced accuracy.
• To improve the performance of DS evidence theory for network anomaly detection.
• To address the challenges posed by complex and varied network features.

Main Methods:

• Implemented an optimized Dempster-Shafer (ODS) evidence theory by assigning weights to sensors based on their prediction accuracy.
• Introduced a regression basic probability assignment (RBPA) function to leverage sensor regression capabilities for complex network analysis.
• Conducted four experimental evaluations to validate the system's effectiveness.

Main Results:

• The novel network anomaly detection system demonstrated a superior detection rate compared to existing methods.
• Both the ODS optimization and RBPA function significantly enhanced overall system performance.
• Experimental results confirmed the efficacy of the proposed approach in identifying network anomalies.

Conclusions:

• The proposed network anomaly detection system offers a significant improvement in detection rates.
• The ODS and RBPA methods are effective in optimizing DS evidence theory and handling network complexity.
• This research provides a robust solution for detecting anomalies in sophisticated computer networks.