Intrusion Detection System Optimization Using ConvXGBoost for Enhanced Threat Detection

Abstract

Enhancing Intrusion Detection Systems (IDS) is critical for strengthening cybersecurity against evolving threats. This research presents a comparative analysis of five machine learning algorithms such as Random Forest (RF), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), XGBoost, and Convolutional XGBoost (ConvXGBoost) for IDS classification. The evaluation is based on key performance metrics, including Accuracy, Precision, Recall, and F1-Score, across multiple attack categories such as DoS, Probe, R2L, and U2R. The experimental results indicate that ConvXGBoost outperforms other models, achieving the highest accuracy (0.97), precision (0.97), recall (0.88), and F1-score (0.93). Furthermore, the integration of Convolutional Neural Networks (CNN) with XGBoost enhances feature extraction, leading to improved classification performance. The research also presents an analysis of training performance over epochs, a confusion matrix for error assessment, and insights into model generalization. The findings highlight the potential of ConvXGBoost in optimizing IDS efficiency, offering a scalable and robust solution for cybersecurity applications.