Abstract
The rapid expansion of networked systems and Internet of Things (IoT) devices has significantly increased the risk of cyber-attacks, making network intrusion detection an essential component of modern cybersecurity. Traditional intrusion detection systems often rely on signature-based tech- niques, which are ineffective against unknown and evolving attacks. This paper proposes an AI-powered intrusion detection system that utilizes machine learning techniques to improve the accuracy and efficiency of attack detection in intelligent network environments. The proposed system employs multiple machine learning algorithms, including Random Forest, AdaBoost, Light Gradient Boosting Machine (LightGBM), and Multi-Layer Per- ceptron (MLP), to classify network traffic as normal or malicious. The system is trained and evaluated using both the NSL- KDD dataset for traditional network intrusions and the Bot-IoT dataset for modern IoT-based attacks, enabling comprehensive analysis across diverse attack scenarios. Experimental results demonstrate that the proposed approach achieves high detection accuracy and improved classification performance compared to traditional methods, while reducing false positive rates. The main contribution of this work lies in the integration of multiple ma- chine learning models and datasets within a unified framework, providing enhanced intrusion detection capability and improved network security for intelligent and dynamic environments.