DEEP LEARNING TECHNIQUES TO ENHANCE ENERGY EFFICIENCY OF HOME APPLIANCES BY ANALYSING AIR QUALITY LEVELS

Abstract

Energy efficiency in home appliances is a critical area of research that addresses the growing demand for reducing energy consumption. The rapid growth in artificial intelligence has prioritized the development of advanced methods to improve sustainable energy consumption, particularly by optimizing the energy efficiency of home appliances. This paper introduces a novel deep learning-based framework to enhance energy efficiency in home appliances by leveraging insights from indoor air quality (IAQ) metrics. Unlike conventional energy management approaches that face challenges such as limited datasets, computational inefficiencies, and lack of generalizability, this research incorporates advanced preprocessing and augmentation techniques. Specifically, a hybrid SMOTE-ENN approach addresses class imbalance, while Z-score normalization ensures consistent feature scaling. Among the evaluated models, Bidirectional GRU and Stacked LSTM stand out, achieving exceptional validation accuracies of 99.81% and 99.64%, respectively, demonstrating superior generalization. This framework uniquely integrates indoor air quality data to optimize energy usage dynamically, showcasing how environmental factors such as CO2, humidity, and temperature can inform sustainable energy practices. These findings underscore the transformative potential of deep learning in fostering eco-friendly innovations for smart home energy management.