Multi-Criteria Feature Selection Based Intrusion Detection for Internet of Things Big Data

Area of Science:

• Computer Science
• Cybersecurity
• Network Security

Background:

• The proliferation of Internet of Things (IoT) devices generates massive amounts of big data, creating significant cybersecurity challenges.
• Real-time network attack detection in resource-constrained IoT environments is difficult due to data complexity and varied traffic feature formats.
• Existing intrusion detection systems struggle with irrelevant features, impacting the speed and accuracy of model training.

Purpose of the Study:

• To develop an effective intrusion detection system for safeguarding IoT big data.
• To address the challenge of irrelevant features in IoT network traffic data that hinder detection accuracy.
• To introduce a novel model that enhances the precision of intrusion detection in IoT environments.

Main Methods:

• The study proposes the LVW-MECO (Las Vegas Wrapper enhanced with multiple evaluation criteria) model.
• LVW-MECO utilizes the Las Vegas Wrapper (LVW) algorithm integrated with multiple evaluation criteria.
• The model is designed to identify pertinent features from IoT network data for improved intrusion detection.

Main Results:

• Experimental results demonstrate the efficacy of LVW-MECO in addressing critical IoT security issues.
• The proposed model significantly enhances intrusion detection performance in IoT networks.
• LVW-MECO effectively improves the accuracy of detecting network attacks by selecting relevant features.

Conclusions:

• LVW-MECO offers a robust solution for enhancing the security of IoT big data.
• The model contributes to safeguarding IoT data integrity and promotes a more secure IoT ecosystem.
• Accurate feature selection is crucial for effective intrusion detection in complex IoT environments.