Forecasting the prices of the day-ahead electricity markets using real data from SEEPEX
Date Issued
2022
Author(s)
Petkovski, Pavle
Dedinec, Aleksandar
Abstract
Forecasting the day-ahead electricity prices can be
significant for every business involved in the electricity market.
In this paper, we compare different machine learning techniques
and algorithms using real data from Serbian Power Exchange,
weather data from Serbian capital city Belgrade and generation
per production type data for Serbian electricity production.
Then on this data, we train different machine learning models:
Linear Regression, Decision Trees, Support Vector Machines,
Random Forest models, Extreme Gradient Boosting models,
Deep Learning models. Metric that we used for comparison
between models is the coefficient of determination.
significant for every business involved in the electricity market.
In this paper, we compare different machine learning techniques
and algorithms using real data from Serbian Power Exchange,
weather data from Serbian capital city Belgrade and generation
per production type data for Serbian electricity production.
Then on this data, we train different machine learning models:
Linear Regression, Decision Trees, Support Vector Machines,
Random Forest models, Extreme Gradient Boosting models,
Deep Learning models. Metric that we used for comparison
between models is the coefficient of determination.
Subjects
File(s)![Thumbnail Image]()
Loading...
Name
CIIT_2022_26.pdf
Size
383.76 KB
Format
Adobe PDF
Checksum
(MD5):8773a536a67291090627331d7bcdca03