An Aggregated Mutual Information Based Feature Selection with Machine Learning Methods for Enhancing IoT Botnet...

Area of Science:

• Cybersecurity
• Machine Learning
• Internet of Things (IoT)

Background:

• Internet of Things (IoT) networks face increasing botnet attacks due to widespread device usage.
• Existing Intrusion Detection Systems (IDS) for IoT struggle with efficiency due to device limitations (battery, processor).
• There is a need for efficient and effective IoT botnet detection methods with low computational time and high accuracy.

Purpose of the Study:

• To propose an aggregated mutual information-based feature selection approach for enhancing the detection of IoT botnet attacks.
• To improve the efficiency and effectiveness of Intrusion Detection Systems (IDS) in resource-constrained IoT environments.
• To evaluate the performance of machine learning classifiers using the proposed feature selection method on real IoT traffic data.

Main Methods:

• Utilized the N-BaIoT benchmark dataset comprising real traffic data from nine commercial IoT devices.
• Employed an aggregated mutual information-based feature selection technique, incorporating Mutual Information (MI), Principal Component Analysis (PCA), and ANOVA f-test.
• Evaluated several ensemble and individual machine learning classifiers, including Random Forest (RF), XGBoost (XGB), Gaussian Naïve Bayes (GNB), k-Nearest Neighbor (k-NN), Logistic Regression (LR), and Support Vector Machine (SVM).

Main Results:

• The proposed aggregated mutual information-based feature selection approach significantly enhanced the detection performance of IoT botnet classifiers.
• The method demonstrated efficiency and effectiveness, outperforming other techniques across various evaluation metrics.
• The feature selection process was applied at a finely-granulated detection level to identify the most relevant features for classification.

Conclusions:

• The developed feature selection approach is effective in improving the accuracy and efficiency of IoT botnet attack detection.
• This method offers a promising solution for building robust and resource-aware Intrusion Detection Systems (IDS) for the Internet of Things (IoT).
• The study highlights the importance of optimized feature selection for machine learning-based cybersecurity in IoT ecosystems.