Πρόγνωση τιμών φυσικού αερίου με Machine Learning =Forecasting Natural Gas Spot Prices with Machine Learning.

Δημήτριος Σάββας Μουχτάρης

Πρόγνωση τιμών φυσικού αερίου με Machine Learning =Forecasting Natural Gas Spot Prices with Machine Learning.

Δημήτριος Σάββας Μουχτάρης

Περίληψη

Η δυνατότητα πρόγνωσης της τιμής του φυσικού αερίου έχει προφανή σημασία για τους παραγωγούς, καταναλωτές και επενδυτές στην σχετική αγορά του φυσικού αερίου. Στην παρούσα διπλωματική γίνεται προσπάθεια πρόγνωσης του natural gas spot price με την χρήση μεθόδων μηχανικής μάθησης και συγκεκριμένα Support Vector Machines (SVM), Regression Trees, Linear Regression, Gaussian Process Regression (GPR) και Ensemble of Trees. Τα μοντέλα εκπαιδεύτηκαν με τη χρήση ενός συνόλου 21 επεξηγηματικών μεταβλητών. Χρησιμοποιούμε την μέθοδο 5 fold-cross-validation με το 90% των δεδομένων για την εκπαίδευση των μοντέλων και το τελευταίο 10% των δεδομένων για τον έλεγχο σε άγνωστα δεδομένα. Τα αποτελέσματα δείχνουν ότι οι μέθοδοι μηχανικής μάθησης παρουσιάζουν διαφορές στην ακρίβεια πρόγνωσης των τιμών του φυσικού αερίου. Ωστόσο τα μοντέλα Bagged Trees (που ανήκει στην μέθοδο Ensemble of Trees) και Linear Regression παρουσιάζουν την καλύτερη απόδοση πρόγνωσης σε σύγκριση με τα υπόλοιπα μοντέλα.

The ability to forecast the price of natural gas benefits stakeholders and has become a valuable tool for all market participants in competitive gas markets. In this paper an attempt is made to forecast the natural gas spot prices using machine learning methods and specifically with Support Vector Machines (SVM), Regression Trees, Linear Regression, Gaussian Process Regression (GPR) and Ensemble of Trees. These models are trained by using a set of 21 explanatory variables. We utilize the 5 fold-cross-validation method with 90% of the natural gas spot price data for model training and the remaining 10% of data for testing purposes of unknown data. The results show that these machine learning methods all have different forecasting accuracy when it comes to forecasting natural gas spot prices. However, the Bagged Trees (belonging to the Ensemble of Trees method) and Linear Regression models have an improved forecasting performance compared to the rest.

Πλήρες Κείμενο:

PDF

Αναφορές

Abdel-Aal, R. E. (2004). Short-Term Hourly Load Forecasting Using Abductive Networks. IEEE TRANSACTIONS ON POWER SYSTEMS, 19(1), 164-173. doi: 10.1109/TPWRS.2003.820695

Aksoy, B., & Selbas, R. (2019, June 14). Estimation of Wind Turbine Energy Production Value by Using Machine Learning Algorithms and Development of Implementation Program. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 1-13. doi:https://doi.org/10.1080/15567036.2019.1631410

Amadeo, K. (2020, September 17). The S&P 500 and How It Works. Retrieved from The Balance: https://www.thebalance.com/what-is-the-sandp-500-3305888

Athanasiou, A. F., Gogas, P., & Papadimitriou, T. (2020, December). Forecasting S&P 500 spikes: an SVM approach. Digital Finance. doi:10.1007/s42521-020-00024-0

Blaskowitz, O., & Herwartz, H. (2011). On economic evaluation of directional forecasts. International Journal of Forecasting, 27(4), 1058-1065. doi:10.1016/j.ijforecast.2010.07.002

Buchanan, W. K., Hodges, P., & Theis, J. (2001, May). Which way the natural gas price: an attempt to predict the direction of natural gas spot price movements using trader positions. Energy Economics, 23(3), 279-293. doi:10.1016/S0140-9883(00)00074-8

Caplinger, D. (2016, July 12). What Are Crude Oil Futures and How Do They Work? Retrieved from TheMotleyFool: https://www.fool.com/investing/2016/07/12/what-are-crude-oil-futures-and-how-do-they-work.aspx#:~:text=Crude%20oil%20futures%20are%20futures,given%20date%20in%20the%20future.

Čeperić, E., Čeperić, V., & Žiković, S. (2017, September 8). Short-term forecasting of natural gas prices using machine learning and feature selection algorithms. (H. Lund, Ed.) Energy, 140, 893-900.

Chen, B. J., Chang, M. W., & Lin, C. J. (2004, November). Load forecasting using support vector Machines: a study on EUNITE competition 2001. IEEE Transactions on Power Systems, 19(4), 1821-1830. doi:10.1109/TPWRS.2004.835679

Chen, J. (2019, June 25). Nasdaq Composite Index. Retrieved from Investopedia: https://www.investopedia.com/terms/n/nasdaqcompositeindex.asp

Chen, J. (2020, January 31). Exchange Rate Definition. Retrieved from Investopedia: https://www.investopedia.com/terms/e/exchangerate.asp

Chen, J. (2020, September 22). Federal Funds Rate. Retrieved from Investopedia : https://www.investopedia.com/terms/f/federalfundsrate.asp

Chen, J. (2020, February 21). One-Year Constant Maturity Treasury. Retrieved from Investopedia: https://www.investopedia.com/terms/c/cmtindex.asp

Chen, J. (2020, June 30). Prime Rate. Retrieved from Investopedia: https://www.investopedia.com/terms/p/primerate.asp

Chen, J. (2020, December 31). West Texas Intermediate (WTI) . Retrieved from Investopedia: https://www.investopedia.com/terms/w/wti.asp

De Felice, M., & Xin Yao. (2011, August). Short-Term Load Forecasting with Neural Network Ensembles: A Comparative Study [Application Notes]. IEEE Computational Intelligence Magazine, 6(3), 47-56. doi:10.1109/MCI.2011.941590

Dimitriadou, A., Gogas, P., Papadimitriou, T., & Plakandaras, V. (2018, October 13). Oil Market Efficiency under a Machine Learning Perspective. Forecasting, 1(1), 157-168. doi:10.3390/forecast1010011

(n.d.). Dow Jones Industrial Average (DJIA). Corporate Finance Institute. Retrieved January 18, 2021, from Corporate Finance Institute: https://corporatefinanceinstitute.com/resources/knowledge/trading-investing/dow-jones-industrial-average-djia/

Drucker, H., Burges, J. C., Kaufman, L., Smola, A., & Vapnik, V. (1996). Support Vector Regression Machines.

Fama, E. F. (1970, May). Efficient Capital Markets: A Review of Theory and Empirical Work. Journal of Finance, 25(2), 383-417.

Interest Rate Spreads. (2021, January 15). Retrieved from FRED: https://fred.stlouisfed.org/series/T5YIE#:~:text=The%20breakeven%20inflation%20rate%20represents,Constant%20Maturity%20Securities%20(TC_5YEAR).

Jordan, M. I., & Mitchell, T. M. (2015, July 17). Science. Machine learning: Trends, perspectives, and prospects, 349(6245), 255-260.

Liu, K., Subbarayan, S., Shoults, S. S., Manry, M. T., Kwan, C., Lewis, F. I., & Naccarino, J. (1996, May). Comparison of very short-term load forecasting techniques. IEEE Transactions on Power Systems, 11(2), 877-882. doi:10.1109/59.496169

Mishra, V., & Smyth, R. (2016, April). Are natural gas spot and futures prices predictable? Economic Modelling, 54, 178-186. doi:10.1016/j.econmod.2015.12.034

Mohandes, M. (2002). Support Vector Machines for short-term electrical load forecasting. Energy Research , 26(4), 335-345. doi:doi.org/10.1002/er.787

Nguyen , H. T., & Nabney, I. T. (2010, September). Short-term electricity demand and gas price forecasts using wavelet transforms and adaptive models. Energy, 35(9), 3674-3685. doi:10.1016/j.energy.2010.05.013

NYMEX: Your Home for Henry Hub Natural Gas. (n.d.). Retrieved from cmegroup: https://www.cmegroup.com/trading/energy/nymex-natural-gas-futures.html

Park, D. C., El-Sharkawi, M. A., Marks, R. J., Atlas, L. E., & Damborg, M. J. (1991, May). Electric load forecasting using an artificial neural network. IEEE Transactions on Power Systems, 6(2), 442-449. doi:10.1109/59.76685

Peterson, K. (2018, September). Resting Heart Rate Variability Can Predict Track and Field Sprint Performance. OA Journal - Sports, 1.

Plakandaras, V., Rangan, G., Gogas, P., & Papadimitriou, T. (2015). Forecasting the U.S. real house price index. Economic Modelling, 45, 259-267.

Rasmussen, C. E. (2004). Gaussian Processes in Machine Learning. In Advanced Lectures on Machine Learning (pp. 63-71).

Salehnia , N., Falahi, M. A., Seifi, A., & Mahdavi Adeli, M. H. (2013, September). Forecasting natural gas spot prices with nonlinear modeling using Gamma test analysis. Journal of Natural Gas Science and Engineering, 14, 238-249. doi:10.1016/j.jngse.2013.07.002

Seeger, M. (2004). Gaussian Processes for Machine Learning . International Journal of Neural Systems, 24(2), 69-106. doi:10.1142/S0129065704001899

Serletis, A., & Shahmoradi, A. (2006, September). Returns and volatility in the NYMEX Henry Hub natural gas futures market. OPEC Energy Review, 30(3). doi:10.1111/j.1468-0076.2006.00167.x

Statistics Solutions. (2013). Retrieved from https://www.statisticssolutions.com/what-is-linear-regression/

Su, M., Zhang, Z., Zhu, Y., Zha, D., & Wen, W. (2019, May 3). Data Driven Natural Gas Spot Price Prediction Models Using Machine Learning Methods. Energies , 12(9). doi:org/10.3390/en12091680

Tan, Y. V., & Roy, J. (2019). Bayesian additive regression trees and the General BART model. In Statistic in Medicine (pp. 5048-5069).

Vapnik, V., Golowich, S. E., & Smola, A. (1996). Support Vector Method for Function Approximation, Regression Estimation and Signal Processing·.

Ying , C., Luh, P., Yige, Z., Michel , L., Coolbeth, M., Guan, C., . . . Rourke, S. J. (2010, February 322-330). Short-Term Load Forecasting: Similar Day-Based Wavelet Neural Networks. IEEE Transactions on Power Systems, 25(1).

Zimmermann, A. (2008). Ensemble-Trees: Leveraging Ensemble Power Inside Decision Trees. In Lecture Notes in Computer Science (Vol. 5255, pp. 76-87). Springer, Berlin, Heidelberg. doi:https://doi.org/10.1007/978-3-540-88411-8_10

Εισερχόμενη Αναφορά

Δεν υπάρχουν προς το παρόν εισερχόμενες αναφορές.