Anomaly Based IDS in Industrial IoT

The Internet of Things (IoT) is a critical component for many industries, and Intrusion Detection Systems (IDS) are implemented to safeguard the smart environment created by IoT devices. This paper aims at comparing the accuracies of different machine learning algorithm used for implementation of Anomaly Based IDS in IoT. A systematic review of recent NIDS techniques has been done which proves that Anomaly-based NIDS is a better detection method. Later the comparative study of different anomaly-based implementation techniques was performed. Comparative study concluded the four machine learning algorithms which are Gaussian Naive Bayes, Decision tree, Random Forest, and Support Vector Machine that provided higher accuracies. Therefore, a Machine learning model for all four algorithms was built and trained on KDDCUP 1999 dataset to compare the accuracies and figure out the best algorithm for AIDS.