Detection of internet of things network attacks by hybrid deep learning (CNN-LSTM) algorithm to enhance security

Area of Science:

• Cybersecurity and Network Engineering
• Artificial Intelligence and Machine Learning

Background:

• The proliferation of Internet of Things (IoT) devices introduces significant cybersecurity vulnerabilities.
• Botnet attacks represent a critical threat to network stability and data integrity in IoT environments.

Purpose of the Study:

• To develop and evaluate a hybrid deep learning model for enhanced botnet attack detection in IoT networks.
• To improve the accuracy and robustness of botnet detection systems using advanced machine learning techniques.

Main Methods:

• A hybrid deep learning model integrating Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) was developed.
• Feature Engineering (FE) and SMOTE were employed to preprocess the BoT-IoT dataset, addressing noise and class imbalance.
• The model was trained and validated on both balanced and imbalanced versions of the BoT-IoT dataset.

Main Results:

• The proposed CNN-LSTM hybrid model achieved exceptional performance, with 99.77% accuracy, 100% PR-AUC, and 99.99% ROC-AUC on a balanced dataset.
• The model demonstrated superior feature representation and temporal pattern recognition compared to traditional deep learning models.
• Effective performance was observed on imbalanced data, highlighting the model's generalizability and robustness.

Conclusions:

• The hybrid CNN-LSTM model offers a scalable and interpretable solution for detecting botnet attacks in IoT ecosystems.
• This approach significantly enhances attack classification accuracy and outperforms existing baseline models.
• The developed model provides a robust defense mechanism against evolving cybersecurity threats in the IoT landscape.