Ensemble-based classification algorithm to enhance stability of energy management in IoT-based smart grid networks

Abstract

The exponential increase in electricity consumption makes renewable energy management a necessity within the global warming context. Internet of things (IoT) has a key role in effective data transmission for better managing of energy dissipation in smart grids. Since smart grid network deployment involves huge complexities due to the large data volume being generated, applying artificial intelligent methods is essential to better manage the process. Moreover, reducing energy consumption in a stable smart grid system and fault detection are important in managing electricity congestions, power failure and grid stability problems. This paper aims to present a novel prediction architecture involving ensemble bagging trees and analysis of variance (ANOVA) as a feature selection strategy to improve stability of energy consumption and maximise prediction factors such as accuracy, precision, recall and F1-score in IoT-based smart grids. Experimental and simulation results show that the proposed architecture can decrease training time and improve accuracy of prediction with 99.999% on validation (training) data and 100% on test data than other state-of-the-are machine learning mechanisms.