A Review of Machine Learning Algorithms Used for Load Forecasting at Microgrid Level

As load forecasting nowadays is a crucial and integral part of the energy production procedures a large number of forecasting methods has been proposed to address it. However, although there are many forecasting methods which take into account the advances in information, metering and control technologies in order to address the challenges of forecasting problems, the accuracy and efficiency levels required for each type of applications are yet to be determined. Technologies such as machine learning techniques have been proven useful for short-term electricity load forecasting especially in microgrids where a large variety of data should be included in the energy consumption prognosis. In this paper, we present an overview of the main machine learning algorithms applied to electricity load datasets for short-term forecasting such as Support Vector Machines (SVM), k-Nearest Neighbors (kNN), Random Forest and Artificial Neural Networks (ANN) and compare their performance efficiency, capabilities and limitations.