A Stacked Ensemble Classifier for an Intrusion Detection System in the Edge of IoT and IIoT Networks

Abstract

Over the last three decades, cyberattacks have become a threat to national security. These attacks can compromise Internet of Things (IoT) and Industrial Internet of Things (IIoT) networks and affect society. In this paper, we explore Artificial Intelligence (AI) techniques with Machine and Deep Learning models to improve the performance of an anomaly-based Intrusion Detection System (IDS). We use the ensemble classifier method to find the best combination between multiple models of prediction algorithms and to stack the output of these individual models to obtain the final prediction of a new and unique model with better precision. Although, there are many ensemble approaches, finding a suitable ensemble configuration for a given dataset is still challenging. We designed an Artificial Neural Network (ANN) with the Adam optimizer to update all model weights based on training data and achieve the best performance. The result shows that it is possible to use a stacked ensemble classifier to achieve good evaluation metrics. For instance, the average accuracy achieved by one of the proposed models was 99.7%. This result was better than the results obtained by any other individual classifier. All the developed code is publicly available to ensure reproducibility.