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

40 Pages Posted: 5 Nov 2020 Last revised: 17 Mar 2021

See all articles by Jiawen Luo

Jiawen Luo

South China University of Technology

Qiang Ji

Chinese Academy of Sciences (CAS) - Institute of Policy and Management

Tony Klein

Queen's University Belfast - Queen's Management School

Thomas Walther

Utrecht University - School of Economics

Date Written: March 15, 2020

Abstract

We augment the HAR model with additional information channels to forecast realized volatility of WTI futures prices. These channels include stock markets, sentiment indices, commodity and FX markets, and text-based Google indices. We then apply four differing machine learning techniques to identify the most suitable endo- and exogenous factors which improve baseline model forecasts. We show that machine learning generated forecasts provide better forecasting quality and that portfolios which are constructed with these forecasts outperform their competing models. We find LASSO and SSVS to provide outperforming forecasts and portfolio weights. Analyzing the selection process, we show that variable choices 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: Forecasting, Crude oil, Realized volatility, Exogenous predictors, Machine learning

JEL Classification: C22, C45, E37, Q47

Suggested Citation

Luo, Jiawen and Ji, Qiang and Klein, Tony and Walther, Thomas, Forecasting Realized Volatility of Crude Oil Futures Prices based on Machine Learning (March 15, 2020). QMS Research Paper 2021/04, Available at SSRN: https://ssrn.com/abstract=3701000 or http://dx.doi.org/10.2139/ssrn.3701000

Jiawen Luo

South China University of Technology ( email )

Wushan
Guangzhou, AR Guangdong 510640
China

Qiang Ji

Chinese Academy of Sciences (CAS) - Institute of Policy and Management ( email )

No.15 ZhongGuanCun BeiYiTiao Alley
Haidian District
Beijing, 100190
China

Tony Klein (Contact Author)

Queen's University Belfast - Queen's Management School ( email )

Riddel Hall
185 Stranmillis Road
Belfast, BT9 5EE
United Kingdom

HOME PAGE: http://www.tony-klein.info

Thomas Walther

Utrecht University - School of Economics ( email )

Kriekenpitplein 21-22
Adam Smith Building
Utrecht, +31 30 253 7373 3584 EC
Netherlands

HOME PAGE: http://www.thomas-walther.info

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
160
Abstract Views
611
rank
219,843
PlumX Metrics