Machine Learning Methods for Energy Sector in IoT

Abstract

The Internet of thing (IoT) field is the subject of many studies today. In general, the IoT is the objects and equipment around us that are connected to the Internet and can be controlled and managed by smartphones and tablets. This concept is used in various fields including: smart homes, smart cities, agriculture, transportation and energy management. Various algorithms for energy management have been proposed in this field. This paper investigates studies in field of machine learning and compares in terms of advantages, disadvantages, implementation environment and learning algorithms. In this paper, we examine the energy efficiency in the Weka tool and use a number of learning algorithms such as Linear Regression, K-Nearest Neighbors, Decision Stump, RBF Network and Isotonic Regression algorithms to predict the Cooling Load and Heating Load in a residential building. The RBF algorithm is equivalent to the neural network algorithm based on the radial basic functions. We also determine the most effective factors on Cooling Load and Heating Load with the help of a feature selection algorithm.