A Mobile Architecture to Manage Residential Electricity Consumption Using IoT-based Smart Plugs and Machine Learning Algorithms

Resumen

This paper proposes a mobile architecture for managing residential electricity consumption data using IoT-based smart plugs and machine learning algorithms. The main objective is to monitor, analyze, and predict electricity consumption in residential environments, aiming to improve energy efficiency and engage users through gamification elements, making energy saving more attractive and motivating. The research addresses these goals through specific questions, hypotheses, and methodological steps, including the analysis of electrical energy consumption data from various household appliances, the development of machine learning algorithms such as Holt-Winters, XGBoost, and Autoencoder LSTM to predict future consumption, and the creation of a prototype mobile application for visualizing and managing residential energy consumption. The Autoencoder LSTM model demonstrated superior accuracy in predicting energy consumption, highlighting its effectiveness. The results underscore the importance of integrating energy consumption prediction technologies and energy management tools in homes to promote sustainability and reduce environmental impact.