Energy Consumption Prediction System for smart homes

ABSTRACT In the era of increasing environmental concerns and rising energy demands, the integration of smart technologies into residential environments has become crucial for achieving sustainability. Smart homes, powered by Internet of Things (IoT) devices and intelligent control systems, offer a promising solution for optimizing energy consumption. This research proposes a machine learning-based predictive model designed to forecast energy usage in smart homes. The model leverages historical consumption data, weather patterns, occupancy behavior, and appliance usage to deliver accurate predictions using algorithms such as Decision Tree, Random Forest, Artificial Neural Networks (ANN), XGBoost, and ensemble learning. The project incorporates a web-based interface using Flask to visualize real-time and predicted consumption data. The results show that ensemble learning techniques achieve a prediction accuracy of up to 99.95%, demonstrating the model’s effectiveness in reducing electricity costs and promoting sustainable living. KEYWORDS Smart Homes, Energy Consumption Prediction, Machine Learning, Sustainability, Time-Series Analysis, Energy Efficiency, Decision Trees, Artificial Neural Networks (ANN)