Machine-Learning Based Energy Consumption Forecasting
Main Article Content
Abstract
Experiment focuses on the development of an accurate energy consumption forecasting system using XGBoost, an advanced and scalable implementation of gradient boosting, to enhance predictive accuracy and efficiency. The methodology integrates data collection, preparation, and exploratory data analysis (EDA), leveraging automated tools for efficient processing. Time series predictions are made using the XGBoost model, optimized through Auto-ML with AutoTS, followed by careful model selection. The chosen model is then deployed using Flask for real-time accessibility. Continuous monitoring and maintenance ensure the model adapts to new data, aiding in effective energy management and cost reduction for manufacturing businesses, while addressing environmental concerns associated with energy consumption.