Prediction of Smart Building Indoor Temperature Using IoT and Machine Learning

Abstract

Due to fast growth and improved living, the need for usable energy has grown astronomically over the last few decades, particularly in the building industry. The energy performance of the building is determined by a number of factors, including the surrounding weather, the structure's features, and the energy consumption pattern. This paper describes a technique for combining the Internet of Things (IoT) with certain widely used machine learning algorithms to develop a predictive model that can be utilized for smart building indoor temperature forecasting. For establishing viability to a completely unknown dataset, this prediction model was developed using an online learning approach. To verify the technique, the article conducts a Machine Learning-based experimental using recorded actual sensor data. Following that, the study recommends incorporating the following technique into an IoT architecture based on Edge Computing to enable the building to operate in an energy-efficient manner.