IoT and Machine Learning Based Prediction of Smart Building Indoor Temperature

The demand for useful energy has increased astronomically over the past few decades, especially in building sector, due to rapid development and enhanced lifestyle. The energy performance of the building is reliant on several parameters like surrounding weather variables, building characteristics and energy usage pattern. This literature highlights a mechanism integrating the Internet of Things (IoT) and some widely used machine Mearning algorithms to create a predictive model that can be used for forecasting of smart building indoor temperature. This predictive model has been trained with on-line learning methodology for developing viability to a completely unfamiliar dataset. The paper carries out a Machine Learning based experimentation on recorded real sensor data [1] to validate the approach. Following that, the paper suggests integration of following strategy into an Edge Computing based IoT architecture for enabling the building to work in an energy-efficient fashion.