AI-Driven Detection of Malicious Network Intrusions Using Decision Tree Algorithm

With the rise of increasingly sophisticated cyber threats, accurate and efficient intrusion detection systems have become essential for securing networks. This paper presents a machine learning-based anomaly detection system using a Decision Tree algorithm, trained on the CTU-IoT-Malware-Capture dataset. The system achieves a high accuracy of $98.82 \%$, outperforming traditional methods in precision and recall. The workflow includes data preprocessing, feature engineering, hyperparameter tuning, and real-time alert integration via SMTP. Compared to alternative models like SVM and Random Forest, this Decision Tree approach offers a balance of interpretability and performance. This research demonstrates the viability of lightweight, interpretable models for real-time cybersecurity solutions, bridging the gap between high detection accuracy and practical deployment in evolving threat landscapes.