TY - JOUR
T1 - Technical indicators for energy market trading
AU - Mc Hugh, Catherine
AU - Coleman, Sonya
AU - Kerr, Dermot
PY - 2021/12/15
Y1 - 2021/12/15
N2 - Technical indicators have been widely applied to the financial trading market, often combined with machine learning algorithms, to predict future stock market prices. The characteristics of energy market data are comparable to financial trading data; hence this research derives eight price prediction technical indicators for hourly electricity prices from the Irish Integrated Single Electricity Market. The proposed indicators consider the three key types of price indicators: trend, oscillator, and momentum. Building the technical indicators from raw electricity price data helps to capture market behaviours and find information to predict future profitable prices. The electricity price data for the proposed indicators were collected from February 2019 until March 2020. Three machine learning regression algorithms were trained with the technical indicators: Extreme Gradient Boosting, Gradient Boosting, and Random Forest. The results demonstrate that the price prediction models perform much better when trained using the proposed technical indicators when compared with baseline raw price data models.
AB - Technical indicators have been widely applied to the financial trading market, often combined with machine learning algorithms, to predict future stock market prices. The characteristics of energy market data are comparable to financial trading data; hence this research derives eight price prediction technical indicators for hourly electricity prices from the Irish Integrated Single Electricity Market. The proposed indicators consider the three key types of price indicators: trend, oscillator, and momentum. Building the technical indicators from raw electricity price data helps to capture market behaviours and find information to predict future profitable prices. The electricity price data for the proposed indicators were collected from February 2019 until March 2020. Three machine learning regression algorithms were trained with the technical indicators: Extreme Gradient Boosting, Gradient Boosting, and Random Forest. The results demonstrate that the price prediction models perform much better when trained using the proposed technical indicators when compared with baseline raw price data models.
KW - Energy market
KW - Hourly price forecasting
KW - Regression machine learning
KW - Technical indicators
UR - https://linkinghub.elsevier.com/retrieve/pii/S2666827021000918
U2 - 10.1016/j.mlwa.2021.100182
DO - 10.1016/j.mlwa.2021.100182
M3 - Article
SN - 2666-8270
VL - 6
SP - 1
EP - 9
JO - Machine Learning with Applications
JF - Machine Learning with Applications
M1 - 100182
ER -