Enhanced intrusion detection system IoT network security model by feed forward neural network and machine learning.

Area of Science:

• Cybersecurity
• Network Security
• Artificial Intelligence

Background:

• Internet of Things (IoT) networks face escalating cyber threats.
• Traditional Intrusion Detection Systems (IDS) are limited by resource constraints and evolving attack patterns.
• Sophisticated attacks in real-time are challenging for current IDS.

Purpose of the Study:

• To develop a novel hybrid IDS integrating deep learning (DL) and machine learning (ML) for improved IoT security.
• To enhance attack detection accuracy while minimizing computational overhead in IoT networks.
• To create a scalable and efficient hybrid IDS model for robust intrusion detection.

Main Methods:

• Proposed a hybrid IDS combining Feed Forward Neural Networks (FFNN) and XGBoost.
• Employed Principal Component Analysis (PCA) for feature selection.
• Trained and evaluated the model on the CIC IoT 2023 dataset for real-time attack data analysis.

Main Results:

• Achieved superior accuracy of 99% in detecting intrusions, surpassing existing IDS techniques.
• The hybrid FFNN-XGBoost model demonstrated better performance than standalone FFNN and XGBoost classifiers.
• Significantly improved precision, recall, and F1-score for robust intrusion detection.

Conclusions:

• The hybrid FFNN-XGBoost IDS offers a scalable and efficient solution for IoT security.
• The study provides valuable insights into addressing challenges like dataset imbalance and feature selection in IDS.
• Findings support future advancements in intrusion detection utilizing DL and ML approaches for IoT environments.