Energy consumption optimization for personalized livability in smart homes using machine learning

Abstract

This project explores a machine learning (ML)-based approach for optimizing energy consumption in smart homes without compromising resident comfort. We propose a system that integrates sensor data analysis, feature engineering, and a Stochastic Gradient Descent Regressor model to predict and manage energy usage. Sensor data encompassing motion, light, temperature, humidity, and current is preprocessed and transformed to extract meaningful features like occupancy patterns and light us age durations. The SGD Regressor model, trained on this data, predicts energy consumption while considering user preferences through a user-friendly web application. The web application empowers residents with real-time monitoring, control functionalities, and personalized recommendations. This user engagement fosters informed energy management while prioritizing comfort. Our results demonstrate the effectiveness of the proposed system in achieving significant energy savings without sacrificing resident livability. This approach paves the way for intelligent smart home systems that personalize energy management while ensuring occupant comfort