Regularization for Electricity Price Forecasting

The most commonly used form of regularization typically involves defining the penalty function as a ℓ1 or ℓ2 norm. However, numerous alternative approaches remain untested in practical applications. In this study, we apply ten different penalty functions to predict electricity prices and evaluate th...

Full description

Saved in:
Bibliographic Details
Main Author: Bartosz Uniejewski
Format: Article
Language:English
Published: Wrocław University of Science and Technology 2024-01-01
Series:Operations Research and Decisions
Online Access:https://ord.pwr.edu.pl/assets/papers_archive/ord2024vol34no3_14.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849430168274206720
author Bartosz Uniejewski
author_facet Bartosz Uniejewski
author_sort Bartosz Uniejewski
collection DOAJ
description The most commonly used form of regularization typically involves defining the penalty function as a ℓ1 or ℓ2 norm. However, numerous alternative approaches remain untested in practical applications. In this study, we apply ten different penalty functions to predict electricity prices and evaluate their performance under two different model structures and in two distinct electricity markets. The study reveals that LQ and elastic net consistently produce more accurate forecasts compared to other regularization types. In particular, they were the only types of penalty functions that consistently produced more accurate forecasts than the most commonly used LASSO. Furthermore, the results suggest that cross-validation outperforms Bayesian information criteria for parameter optimization, and performs as well as models with ex-post parameter selection. (original abstract)
format Article
id doaj-art-427fbe719d334ce89b84716dc7bfa09f
institution Kabale University
issn 2081-8858
2391-6060
language English
publishDate 2024-01-01
publisher Wrocław University of Science and Technology
record_format Article
series Operations Research and Decisions
spelling doaj-art-427fbe719d334ce89b84716dc7bfa09f2025-08-20T03:28:06ZengWrocław University of Science and TechnologyOperations Research and Decisions2081-88582391-60602024-01-01vol. 34no. 3267286171700392Regularization for Electricity Price ForecastingBartosz Uniejewski0Wroclaw University of Science and Technology, Wroclaw, PolandThe most commonly used form of regularization typically involves defining the penalty function as a ℓ1 or ℓ2 norm. However, numerous alternative approaches remain untested in practical applications. In this study, we apply ten different penalty functions to predict electricity prices and evaluate their performance under two different model structures and in two distinct electricity markets. The study reveals that LQ and elastic net consistently produce more accurate forecasts compared to other regularization types. In particular, they were the only types of penalty functions that consistently produced more accurate forecasts than the most commonly used LASSO. Furthermore, the results suggest that cross-validation outperforms Bayesian information criteria for parameter optimization, and performs as well as models with ex-post parameter selection. (original abstract)https://ord.pwr.edu.pl/assets/papers_archive/ord2024vol34no3_14.pdf
spellingShingle Bartosz Uniejewski
Regularization for Electricity Price Forecasting
Operations Research and Decisions
title Regularization for Electricity Price Forecasting
title_full Regularization for Electricity Price Forecasting
title_fullStr Regularization for Electricity Price Forecasting
title_full_unstemmed Regularization for Electricity Price Forecasting
title_short Regularization for Electricity Price Forecasting
title_sort regularization for electricity price forecasting
url https://ord.pwr.edu.pl/assets/papers_archive/ord2024vol34no3_14.pdf
work_keys_str_mv AT bartoszuniejewski regularizationforelectricitypriceforecasting