Intrusion detection by machine learning for multimedia platform.

Area of Science:

• Cybersecurity
• Machine Learning
• Network Security

Background:

• Multimedia platforms like Netflix saw subscriber growth during the pandemic.
• Network attacks pose a threat to multimedia service availability and security.
• Intrusion detection systems (IDS) are crucial for protecting these platforms.

Purpose of the Study:

• To propose an intelligent intrusion detection system for the IP Multimedia Subsystem (IMS).
• To enhance the accuracy of intrusion detection through critical feature selection.
• To evaluate the effectiveness of machine learning classifiers for network attack detection.

Main Methods:

• Developed an intelligent IDS for IMS security using machine learning.
• Employed two-class classifiers: Decision Tree, Support Vector Machine, and Naive Bayesian.
• Selected six critical features (Service, dst_host_same_srv_rate, Flag, Protocol Type, Dst_host_rerror_rate, Count) based on feature ranking.

Main Results:

• The selected critical features demonstrably improved intrusion detection accuracy.
• The proposed method achieved better accuracy compared to state-of-the-art deep learning approaches.
• Feature selection proved vital for optimizing classifier performance in intrusion detection.

Conclusions:

• Intelligent feature selection is key to developing accurate intrusion detection systems for multimedia platforms.
• Machine learning-based IDS offers a robust solution against network attacks in IMS.
• The proposed method provides a more accurate and efficient alternative to existing deep learning techniques.