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

In this study, a machine learning technique is proposed to forecast energy usage, aiming to enhance efficiency across different settings. Data utilized in this research was obtained from Kaggle and comprises variables such as ambient temperature, humidity levels, building area, timestamps, and operational details of energy-related systems like lighting and HVAC. Several predictive algorithms-including XGBoost, multilayer perceptron regression, Ridge, and ElasticNet-were trained and fine-tuned via grid search to maximize forecasting precision. XGBoost emerged as the most effective model, delivering higher accuracy and performance compared to the other algorithms. Comparative analysis with earlier research underscores the notable advancements introduced by this method. Furthermore, benchmarking against established techniques in existing literature reinforces the reliability and strong performance of the proposed approach. Leveraging environmental and operational data to forecast energy usage, this research introduces a viable solution for enhancing energy efficiency and minimizing operational costs within intelligent energy infrastructures. The findings highlight the impactful role of cutting-edge machine learning techniques in advancing modern energy optimization practices.