Electrical Devices Identification Driven by Features and Based on Machine Learning

The analysis of energy data of electrical devices in Smart Homes (SHs) represents an important factor for the decision-making process of energy management both from the consumer perspective by saving money and also in terms of energy redistribution and CO\(_{2}\) emissions reduction, by knowing how the energy demand of a building is composed in the Smart Grid (SG). A proactive monitoring and control mechanism motivates the need to face with the identification of appliances. In this context, the paper proposes a model for the automatic identification of electrical devices driven by 19 features that are formalized through a mathematical notation. On the basis of such proposed features, three different classifiers are trained and experimented, by evaluating their accuracy, for the identification of 33 types of appliances.