Ensemble-Based Approach for Effective Detection of Injection Attacks

Due to the growing number of connected devices and the sophistication of attack vectors, cybersecurity experts are increasingly concerned about identifying injection attacks in Internet of Things (IoT) environments. This work focuses on exploiting the IoT-23 dataset, an exhaustive collection of IoT network traffic, to improve the detection of injection attacks through the application of ensemble learning models. the use of four cutting-edge ensemble techniques: AdaBoost, Light Gradient Boosting Machine (LightGBM), eXtreme Gradient Boosting (XGBoost), and Gradient Boosting Machine (GBM). After being trained separately, the models are integrated to provide a strong detection framework. When compared to individual classifiers, the experimental findings show that the ensemble technology greatly enhances detection accuracy, precision, recall, and F1-score. We shed light on the traits of injection attacks and the essential features for identifying them by examining the feature importance and each model’s contribution inside the ensemble. The ensemble model improves the detection system’s robustness and generalizability in a variety of IoT environments in addition to displaying higher performance metrics. This study demonstrates how ensemble learning can strengthen IoT security and pave the way for future developments in the identification and defence against sophisticated cyber threats.