A Comparative Analysis of Supervised Machine Learning Algorithms for Electricity Demand Forecasting

Rawal, Keerti; Ahmad, Aijaz

doi:10.1109/icpc2t53885.2022.9776960

Cited by 7 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The RFR method demonstrates robustness and superior speed and prevents overfitting. The integration of randomized decision trees leads to a decrease in prediction variance and a decrease in generalization error [65].…”

Section: Random Forest Regression Modelmentioning

confidence: 99%

“…The principal concept underlying boosting is to aggregate a collection of decision trees through an iterative process to generate a robust learner [25], [63]. GB builds a model by adding stages of weak prediction algorithms and optimizing loss functions [65]. In every stage, a regression tree is fitted to the negative gradient of the provided loss function [68].…”

Section: Gradient Boosting Regression Modelmentioning

confidence: 99%

See 1 more Smart Citation

Predicting Energy Demand Using Machine Learning: Exploring Temporal and Weather-Related Patterns, Variations, and Impacts

Sahin,

Udeh,

Wanik

et al. 2024

IEEE Access

View full text Add to dashboard Cite

This study aims to develop models for predicting hourly energy demand in the State of Connecticut, USA from 2011 to 2021 using machine learning algorithms inputted with airport weather stations' data from the Automated Surface Observing System (ASOS), demand data from ISO New England (ISO-NE). We built and evaluated nine different model experiments for each machine learning algorithm for each hour of the day addressing energy demand patterns, variations between workdays and weekends, and COVID-19 impacts. Error metrics analysis results highlighted the GBR model demonstrated better performance compared to the MPR and RFR models. Incorporating both temporal and weather features in the models resulted in a noticeable improvement in error metrics. A consistent overestimation trend was observed for all models during the validation period (2018-2019) which may be attributed to energy efficiency measures and integration of behind-the-meter generation, with a further notable increase in overestimation following the onset of COVID-19 due to a change of habits during the pandemic in addition to decarbonization initiatives in the State. This study emphasizes the need for adapting models to dynamic consumption and weather patterns for improved grid management.

show abstract

Section: Random Forest Regression Modelmentioning

confidence: 99%

Section: Gradient Boosting Regression Modelmentioning

confidence: 99%

Predicting Energy Demand Using Machine Learning: Exploring Temporal and Weather-Related Patterns, Variations, and Impacts

Sahin,

Udeh,

Wanik

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…In [22], the authors applied two ensemble learning methods, the XGBoost regressor and RFR, to forecast demand for the next day during the pandemic period. In the work [23], the authors employed machine learning methods, including linear regression (LR), multivariate polynomial regression, SVR, gradient boosting regressor (GBR), RFR, and K-neighbors regressor, to predict energy demand in New South Wales, Australia. In [24], the authors developed a clustering-based method for electricity prediction that was evaluated using a dataset with data from 105 substations.…”

Section: Related Workmentioning

confidence: 99%

A Demand Forecasting Strategy Based on a Retrofit Architecture for Remote Monitoring of Legacy Building Circuits

et al. 2023

View full text Add to dashboard Cite

Energy demand forecasting is crucial for planning and optimizing the use of energy resources in building facilities. However, integrating digital solutions and learning techniques into legacy buildings presents significant challenges due to limited or outdated resources, hampering predictive analytics in these buildings and their circuits. To fill this gap, this article proposes an innovative demand forecasting strategy using an AIoT retrofit architecture based on the SmartLVGrid metamodel. This architecture allows remote monitoring of legacy building circuits, facilitating the collection, processing and storage of data in the cloud. We use several learning algorithms, including linear regression, support vector regressor, random forest regressor, XGBoost regressor, and long short-term memory (LSTM) neural network, to predict energy demand 15 min ahead, identifying potential overruns of contracted demand in accordance with Brazilian regulations. After Bayesian optimization, the LSTM neural network outperformed other models for most of the selected datasets and detected 32 out of 38 demand overruns on the test set. XGBoost and random forest followed closely, detecting 30 demand overruns. Overall, our cost-effective solution optimizes energy usage and efficiently mitigates potential demand exceedances in building installations. This is achieved through a step-by-step approach to upgrading existing aging facilities, which promotes energy efficiency and sustainability.

show abstract

Hybrid LSTM-Graph Convolutional Neural Network with Wavelet Transform and Correlation Analysis for Electrical Demand Forecasting

Rawal,

Ahmad

2024

SN COMPUT. SCI.

View full text Add to dashboard Cite

A Comparative Analysis of Supervised Machine Learning Algorithms for Electricity Demand Forecasting

Cited by 7 publications

References 16 publications

Predicting Energy Demand Using Machine Learning: Exploring Temporal and Weather-Related Patterns, Variations, and Impacts

Predicting Energy Demand Using Machine Learning: Exploring Temporal and Weather-Related Patterns, Variations, and Impacts

A Demand Forecasting Strategy Based on a Retrofit Architecture for Remote Monitoring of Legacy Building Circuits

Hybrid LSTM-Graph Convolutional Neural Network with Wavelet Transform and Correlation Analysis for Electrical Demand Forecasting

Contact Info

Product

Resources

About