Hierarchical classification method of electricity consumption behaviour of power users based on combination model

Abstract

The field of computer architecture has long studied electricity consumption in depth. While energy acquisition as a machine learning metric is gaining traction, most experiments are still focused on achieving extremely high levels of accuracy with no computational constraints. This study discussed a novel technique for analysing power grid electricity consumption using a classification model and deep learning techniques. The dataset was collected and processed as part of a power grid-based electricity analysis. The processed data were then classified using a hierarchical combination model comprised of photovoltaic cell integrated spatio Gaussian markov convolutional neural networks. The experimental analysis included energy efficiency, power analysis, accuracy, mean absolute percentage error (MAPE), and root mean square error (RMSE). After 10 epochs, deep learning also demonstrated competitive performance with accuracy in training and testing. Machine learning models could be useful in comparing firm energy usage and identifying areas for energy efficiency improvement. The proposed technique achieved 95% energy efficiency, 83% power analysis, 93% accuracy, 81% MAPE, 65% RMSE, and 93% specificity.