Northwestern European wholesale natural gas prices: comparison of several parametric and non-parametric forecasting methods Online publication date: Mon, 10-Aug-2020
by Hassan Hamie; Anis Hoayek; Michel Kamel; Hans Auer
International Journal of Global Energy Issues (IJGEI), Vol. 42, No. 3/4, 2020
Abstract: The ability to understand the stochastic process that governs the changes in natural gas prices is crucial for many reasons. This paper aims to introduce the important methods widely used in econometrics, by linking them to a common 'use case' in the subject of commodity pricing. Using the data of natural gas' weekly prices from 2007 to 2014 of the German gas hub, the methods of least squares, maximum likelihood, machine learning gradient descent, and least squares optimisation are used to compute the coefficients of a multivariate causal regression analysis. This study also tests the short-term prediction of wholesale natural gas prices for each method used. It is found that where the linear approximation is not valid, the method suffers accordingly. However, the mathematical methods of gradient descent and least squares optimisation help visualise the data sets, highlight, and accentuate the nonlinear effects of several variables on the spot gas prices.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Global Energy Issues (IJGEI):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com
Our Blog 
Follow us on Twitter 
Visit us on Facebook