GreenTrust: Trust Assessment Using Ensemble Learning in Internet of Microgrid Things

Abstract

With the rise in industries and population, electricity demand is increasing daily. Microgrids play a crucial role in providing green energy by utilizing renewable energy resources. Microgrids not only help meet the growing electricity demand but also reduce global warming and greenhouse effects. However, many homeowners are hesitant or reluctant to share their excess energy resources with other Microgrid or traditional electric grid users. In this article, we propose a hybrid deep learning and machine learning stacking model named GreenTrust. GreenTrust consists of three evaluation deep learning models at the base level and a single machine learning model at the meta-level. GreenTrust first establishes trust among home users using trust parameters. Once trust is buildup, a Microgrid can share its resources with other grid users. Results show that the hybrid model outperforms than other standalone machine learning schemes, such as Random Forest, XGBoost, and AdaBoost, in terms of accuracy, precision, recall, and F1 score.