부산대학교 도서관

In home area networks, many appliances share a power distribution network and all are potentially the cause and victims of sudden current, voltage, and power spikes. This article proposes a monitoring framework to protect the devices and the network against damage and to optimize power consumption. The authors study and evaluate two machine learning algorithms, support vector machine and k-means clustering, for identifying anomalies and misbehavior, and find that support vector machines seem to be better suited for this application.