A KNN-Based Intrusion Detection System for Enhanced Anomaly Detection in Industrial IoT Networks

Abstract

Anomaly detection in industrial Internet of Things (IIoT) networks is critical for ensuring robust security against evolving cyber threats. Traditional Intrusion Detection Systems (IDS) face challenges such as imbalanced data, sparse anomalies, and high false positive rates, which hinder their effectiveness. This study addresses these limitations by proposing a machine learning-based IDS leveraging the K-Nearest Neighbors (KNN) algorithm. The system efficiently identifies and classifies network anomalies by preprocessing large-scale data streams, selecting relevant features, and optimizing classification tasks. The proposed system demonstrates high precision (87.17%), recall (94.23%), F1-score (90.56%), and accuracy (92.76%) using publicly available intrusion detection datasets. These metrics indicate a significant improvement over existing methods, showcasing the model’s ability to adapt to varying network conditions and detect anomalies with minimal false alarms. This research contributes to enhancing IIoT security by providing a scalable, efficient, and reliable solution for real-time anomaly detection. The findings underline the importance of integrating advanced machine learning techniques into IDS frameworks, paving the way for more secure industrial environments. Future work may explore the integration of deep learning models and real-time deployment in dynamic IIoT ecosystems.