Enhanced Anomaly Detection in Wireless 5G Networks With Hybrid Learning Technique Using AWID3 Dataset

In recent years, the expansion of the Internet of Things and 5G networks has significantly increased wireless traffic, heightening the risk of cyberattacks. Intrusion detection systems have become essential for safeguarding wireless networks by providing real-time threat detection and response. This study presents a comprehensive review and implementation of machine learning-based techniques for detecting various types of wireless attacks, with a focus on improving detection accuracy through ensemble learning. The AWID3 dataset, based on the IEEE 802.11 standard, was used for experimentation. The study was conducted in multiple phases: (1) evaluating six machine learning algorithms (random forest, J48, naïve Bayes, logistic regression, decision tree, and deep neural networks) using three feature selection methods (information gain, gain ratio, and chi-squared); (2) developing a hybrid ensemble model by integrating the strengths of deep neural network, random forest, XGBoost, and LightGBM, with logistic regression as a meta-classifier; and (3) validating performance using key metrics: accuracy, precision, recall, and F1-score. The proposed hybrid model achieved a peak accuracy of 99.75%, outperforming benchmark models in the literature. These results demonstrate the superior performance and robustness of the proposed hybrid approach. By addressing multiple network layers and leveraging ensemble learning, this research highlights the critical role of hybrid models in achieving reliable and accurate intrusion detection for wireless environments.