Estimation of global natural gas spot prices using big data and symbolic regression

被引：0
|
作者：
Stajić L. [1 ]
Praksová R. [2 ]
Brkić D. [2 ,3 ]
Praks P. [2 ]
机构：
[1] Alfatec, Niš
[2] IT4Innovations, VSB – Technical University of Ostrava, Ostrava
[3] Faculty of Electronic Engineering, University of Niš, Niš
来源：
Resources Policy | 2024年 / 95卷
基金：
欧盟地平线“2020”;
关键词：
Market price forecasting; Natural gas; Sensitivity analysis; Spot prices; Supercomputing; Symbolic regression;
D O I：
10.1016/j.resourpol.2024.105144
中图分类号：
学科分类号：
摘要：
This article provides an estimation of future natural gas spot prices on the global international market based on symbolic regression where the sensitivity analysis is performed to identify the most important input parameters. Numerical data sets, comprising various parameters, some of which demonstrate stronger correlations with the global spot price of natural gas, are utilised in this context. PySR (Python Symbolic Regression), a free and open-source software for symbolic regression written in Python, and Julia is used for the presented analysis. Based on the accuracy of the prediction and after sensitivity analysis performed in SALib software, some of the parameters are discovered to be more influencing on natural gas prices compared with others, making this approach suitable for further deeper energy analysis. The analysis shows that in general, global prices of natural gas are influenced mostly by the price of crude oil. The article also presents an overview of methods for predicting natural gas prices with a complementary contribution (interpretable models provided by symbolic regression and sensitivity analysis) tested on the real gas price time-series dataset. © 2024 Elsevier Ltd
引用
收藏
相关论文
共 50 条
  • [31] Application of fuzzy regression analysis using the example of estimation of apartment prices in warsaw
    Minasowicz, Andrew
    9TH INTERNATIONAL CONFERENCE: MODERN BUILDING MATERIALS, STRUCTURES AND TECHNIQUES, VOLS 1-3, 2008, : 312 - 316
  • [32] Prediction with Partitioning: Big Data Analytics Using Regression Techniques
    Saritha, K.
    Abraham, Sajimon
    2017 INTERNATIONAL CONFERENCE ON NETWORKS & ADVANCES IN COMPUTATIONAL TECHNOLOGIES (NETACT), 2017, : 208 - 214
  • [33] State and Parameter Estimation for Natural Gas Pipeline Networks Using Transient State Data
    Sundar, Kaarthik
    Zlotnik, Anatoly
    IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, 2019, 27 (05) : 2110 - 2124
  • [34] Big Geospatial Data Analytics for Global Mangrove Biomass and Carbon Estimation
    Tang, Wenwu
    Zheng, Minrui
    Zhao, Xiang
    Shi, Jiyang
    Yang, Jianxin
    Trettin, Carl C.
    SUSTAINABILITY, 2018, 10 (02)
  • [35] A generalized Schwartz model for energy spot prices - Estimation using a particle MCMC method
    Brix, Anne Floor
    Lunde, Asger
    Wei, Wei
    ENERGY ECONOMICS, 2018, 72 : 560 - 582
  • [36] Nonlinear maximum likelihood estimation of electricity spot prices using recurrent neural networks
    Derrick Mirikitani
    Nikolay Nikolaev
    Neural Computing and Applications, 2011, 20 : 79 - 89
  • [37] Nonlinear maximum likelihood estimation of electricity spot prices using recurrent neural networks
    Mirikitani, Derrick
    Nikolaev, Nikolay
    NEURAL COMPUTING & APPLICATIONS, 2011, 20 (01): : 79 - 89
  • [38] Future spot gas prices in the US and the UK: are movements more influenced by country factors or by global factors?
    Simpson, John
    JOURNAL OF ENERGY MARKETS, 2011, 4 (03) : 3 - 16
  • [39] Opportunities and challenges of using big data for global health
    Peng Jia
    Hong Xue
    Shiyong Liu
    Hao Wang
    Lijian Yang
    Therese Hesketh
    Lu Ma
    Hongwei Cai
    Xin Liu
    Yaogang Wang
    Youfa Wang
    Science Bulletin, 2019, 64 (22) : 1652 - 1654
  • [40] Designing global manufacturing networks using Big Data
    Goelzer, Philipp
    Simon, Lothar
    Cato, Patrick
    Amberg, Michael
    9TH CIRP CONFERENCE ON INTELLIGENT COMPUTATION IN MANUFACTURING ENGINEERING - CIRP ICME '14, 2015, 33 : 191 - 196
12345