Anomaly Detection using ML in Real-Time Systems

Abstract

Modern day real-time systems have been applied in many domains like finance, healthcare, security, and industries, among others. Real-time systems are characterized by continuous creation of huge amounts of data, thus making manual detection of anomalies difficult. Anomaly detection is very essential as anomalies can be used to detect errors in real-time systems and even cyber attacks. In this study, an anomaly detection system using machine learning methods is presented. The proposed model adopts several methods like data preprocessing, feature extraction, supervised and unsupervised machine learning methods to efficiently detect anomalies. The system is able to analyze streams of data, therefore, providing fast and accurate detection of anomalies with minimal time delays.