Enhanced Intrusion Detection System using Decision Tree Classifier for Network Attack Identification and Security Improvement

The escalating frequency and sophistication of network attacks, such as Distributed Denial of Service (DDoS) and port scanning, pose significant threats to data security and system availability. Existing intrusion detection systems (IDS) often fall short in terms of accuracy, real-time detection capabilities, and scalability. This paper presents an enhanced Intrusion Detection System (IDS) utilizing a Decision Tree Classifier to accurately identify and classify network traffic. Leveraging the CICIDS dataset, the proposed system integrates robust preprocessing techniques, including data normalization and Synthetic Minority Oversampling Technique (SMOTE), to address class imbalance. The model's performance is evaluated using standard metrics such as accuracy, precision, recall, and F1-score, achieving competitive results while maintaining interpretability and computational efficiency. Furthermore, a user-friendly web interface, developed using Streamlit, ensures real-time monitoring and ease of deployment. The system demonstrates its potential as a scalable, cost-effective, and efficient solution for dynamic network environments, significantly enhancing security measures against contemporary network threats.