Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning

Jiawen Luo; Tony Klein; Thomas Walther; Qiang Ji

doi:10.1002/for.3077

Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning

Jiawen Luo
, Tony Klein
, Thomas Walther^*
, Qiang Ji

^*Corresponding author for this work

Research output: Contribution to journal › Article › Academic › peer-review

Abstract

Extending the popular HAR model with additional information channels to forecast realized volatility of WTI futures prices, we show that machine learning-generated forecasts provide better forecasting quality and that portfolios that are constructed with these forecasts outperform their competing models resulting in economic gains. Analyzing the selection process, we show that information channels vary across forecasting horizon. Variable selection produces clusters and provides evidence that there are structural changes with regard to the significance of information channels.

Original language	English
Pages (from-to)	1422-1446
Number of pages	25
Journal	Journal of Forecasting
Volume	43
Issue number	5
Early online date	19 Feb 2024
DOIs	https://doi.org/10.1002/for.3077
Publication status	Published - Aug 2024

Bibliographical note

Publisher Copyright:
© 2024 The Authors. Journal of Forecasting published by John Wiley & Sons Ltd.

Funding

We are thankful to Derek Bunn (editor) and an anonymous referee for their comments. An earlier version of this working paper was circulated under the title “Forecasting Realized Volatility of Crude Oil Futures Prices based on Variable Selection Approaches.” Jiawen Luo is supported by the National Natural Foundation of China (grant number ofs 72171088 and 71803049), the Natural Foundation of Guangdong Province (grant number 2023A1515012527), the Guangdong Philosophy and Social Science Foundation (grant number GD23CGL01), the Guangzhou Municipal Science and Technology Bureau (grant number 2023A04J1298).

Funders	Funder number
National Natural Science Foundation of China	72171088, 71803049
Natural Science Foundation of Guangdong Province	2023A1515012527
Philosophy and Social Science Foundation of Hunan Province	GD23CGL01
Guangzhou Municipal Science and Technology Bureau	2023A04J1298

Keywords

crude oil
exogenous predictors
forecasting
machine learning
realized volatility

Access to Document

10.1002/for.3077Licence: CC BY-NC

Journal of Forecasting - 2024 - Luo - Forecasting realized volatility of crude oil futures prices based on machine learningFinal published version, 3.37 MBLicence: CC BY-NC

Cite this

@article{557a7e6c99f543b1a857fec86e27a078,

title = "Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning",

abstract = "Extending the popular HAR model with additional information channels to forecast realized volatility of WTI futures prices, we show that machine learning-generated forecasts provide better forecasting quality and that portfolios that are constructed with these forecasts outperform their competing models resulting in economic gains. Analyzing the selection process, we show that information channels vary across forecasting horizon. Variable selection produces clusters and provides evidence that there are structural changes with regard to the significance of information channels.",

keywords = "crude oil, exogenous predictors, forecasting, machine learning, realized volatility",

author = "Jiawen Luo and Tony Klein and Thomas Walther and Qiang Ji",

note = "Publisher Copyright: {\textcopyright} 2024 The Authors. Journal of Forecasting published by John Wiley \& Sons Ltd.",

year = "2024",

month = aug,

doi = "10.1002/for.3077",

language = "English",

volume = "43",

pages = "1422--1446",

journal = "Journal of Forecasting",

issn = "0277-6693",

publisher = "John Wiley and Sons Ltd",

number = "5",

}

TY - JOUR

T1 - Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning

AU - Luo, Jiawen

AU - Klein, Tony

AU - Walther, Thomas

AU - Ji, Qiang

PY - 2024/8

Y1 - 2024/8

N2 - Extending the popular HAR model with additional information channels to forecast realized volatility of WTI futures prices, we show that machine learning-generated forecasts provide better forecasting quality and that portfolios that are constructed with these forecasts outperform their competing models resulting in economic gains. Analyzing the selection process, we show that information channels vary across forecasting horizon. Variable selection produces clusters and provides evidence that there are structural changes with regard to the significance of information channels.

AB - Extending the popular HAR model with additional information channels to forecast realized volatility of WTI futures prices, we show that machine learning-generated forecasts provide better forecasting quality and that portfolios that are constructed with these forecasts outperform their competing models resulting in economic gains. Analyzing the selection process, we show that information channels vary across forecasting horizon. Variable selection produces clusters and provides evidence that there are structural changes with regard to the significance of information channels.

KW - crude oil

KW - exogenous predictors

KW - forecasting

KW - machine learning

KW - realized volatility

UR - https://www.scopus.com/pages/publications/85185664082

U2 - 10.1002/for.3077

DO - 10.1002/for.3077

M3 - Article

SN - 0277-6693

VL - 43

SP - 1422

EP - 1446

JO - Journal of Forecasting

JF - Journal of Forecasting

IS - 5

ER -

Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning

Abstract

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Forecasting Realized Volatility of Crude Oil Futures Prices based on Variable Selection Approaches

Cite this