Application of Neighborhood Component Analysis for Enhancing Load Recognition in Home Energy Management Systems

Thales Wulfert Cabral, Fernando Bauer, Eduardo de Lima, Gustavo Fraidenraich, Luís Geraldo P. Meloni

DOI: 10.14209/sbrt.2024.1571032518

Evento: XLII Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2024)

Keywords: Neighborhood Component Analysis NCA Machine learning Load Recognition

Abstract

The increasing residential demand for electricity has a direct effect on the balance between human activities and the environment. Technological solutions like the Home Energy Management System (HEMS) are essential for sustainable energy consumption. This paper proposes novel approaches to load recognition in HEMS, which main contributions are enhanced performance through the application of Neighborhood Component Analysis (NCA) jointly with (i) the optimized Support Vector Machine (SVM), (ii) the optimized k-nearest Neighbors (k-NN), (iii) the optimized Extreme Gradient Boosting (XGBoost), and (iv) improved results for training and inference times.

Download