Applying classification methods to model standby power consumption in the Internet of Things

Abstract

This paper presents an approach that combines Internet of Things (IoT) technologies and classification methods to improve efficient usage of power consumption. We focused on energy use of electronic devices on standby mode, which represent from 5 to 26% of power consumption in a home. The proposed approach aims at predicting situation in which devices on standby can be turned off, reducing power consumption. In summary, our approach uses motion and current sensors connected to an IoT infrastructure to build a profile about the presence of people at home. Results obtained from our approach present a reduction of the electric energy consumption by applying Machine Learning methods on Internet of Things scenarios.