Using spatiotemporal patterns, this letter provides a flexible machine learning detection strategy for cyberattacks in distribution systems. Based on system-wide observations, the graph Laplacian detects spatiotemporal patterns. When cyberattacks happen, a flexible Bayes classifier (BC) teaches spatiotemporal practices, resulting in a violation. Flexible BCs used online can potentially be used to detect cyberattacks. Through the use of common IEEE 13- and 123-node test feeds, the effectiveness of the devised approach is shown. In order to anticipate load under cyberattacks, this research presents a machine learning-based anomaly detection (MLAD) technique

Cyber attack detection, distribution systems, graph Laplacian, machine learning, spatiotemporal