Publication Date: 2023/09/15
Abstract: The only byproducts of burning natural gas are carbon dioxide, water vapor, and very little amounts of nitrogen oxide, making it the cleanest fossil fuel on the planet. A wide range of consumer products, such as stoves, dryers, fireplaces, and furnaces, are also powered by natural gas. At least one of your appliances undoubtedly runs on natural gas. In this work, the demand for residential natural gas was forecasted using a hybrid ensemble regression machine learning approach. Accurate forecasting of the demand for natural gas is crucial for effective energy management and resource allocation. The hybrid ensemble approach mixes a number of regression algorithms, including linear regression (LR), decision tree regression (DTR), support vector regression (SVR), and K-nearest neighbor (KNN), to take advantage of the benefits of each unique model and improve prediction performance. The hybrid ensemble regression model's process has two steps. In the first stage, distinct regression models are trained on the dataset. The second stage involves evaluating each model's predictions. To evaluate the effectiveness of the hybrid ensemble model, a range of measures, including mean absolute error (MAE), mean squared error (MSE), coefficient of determination (R-squared), and accuracy, are generated and compared to those of individual regression models. The anticipated accuracy of the model is further assessed using cross-validation techniques to ensure resilience. The results of the experiment demonstrated that the hybrid ensemble regression technique routinely outperformed individual regression models in terms of prediction accuracy. Combining numerous models enables the collection of the various correlations and patterns contained in the data, enhancing the model's overall performance.
Keywords: Ensemble, Hybrid, Machine Learning, Natural Gas, Prediction.
DOI: https://doi.org/10.5281/zenodo.8348415
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23SEP126.pdf
REFERENCES