کنفرانس بین المللی فناوری و مدیریت انرژی

صفحه اصلی / دهمین کنفرانس بین المللی فناوری و مدیریت انرژی

Data-Driven Energy Consumption Prediction: A Comprehensive Approach to Smart Energy Management

نویسندگان :

Soheil Sheikh Ahmadi¹ Alireza Sheikh Ahmadi²

1- دانشگاه تبریز 2- دانشگاه ارومیه

کلمات کلیدی :

Energy Consumption Prediction،Machine Learning،MLP Regression Optimization،Smart Energy Systems،XGBoost

چکیده :

This paper presents a machine learning-based approach for predicting energy consumption to optimize energy usage in various environments. The dataset, sourced from Kaggle, includes features like temperature, humidity, square footage, time of day, and operational statuses of energy systems such as HVAC and lighting. Multiple models, including XGBoost, MLP regression, Ridge, Lasso, and ElasticNet, were trained and optimized using grid search to enhance prediction accuracy. Among these, XGBoost outperformed other methods, demonstrating superior accuracy and efficiency. A detailed comparison with previous studies further highlights significant improvements achieved by this approach. Additionally, the proposed method has been evaluated against other techniques reported in the literature, confirming its robustness and effectiveness. By accurately predicting energy consumption through environmental and operational parameters, this study offers a practical tool for efficient energy management and cost reduction in smart energy systems. The results underscore the potential of advanced machine learning models for improving energy optimization strategies.