Academic journal article Asian Social Science

Short-Term Fuzzy Forecasting of Brent Oil Prices

Academic journal article Asian Social Science

Short-Term Fuzzy Forecasting of Brent Oil Prices

Article excerpt


Oil prices movements is very important macroeconomic factor for decision making. The accuracy of results for different types of oil brands depends on models and algorithms. This paper evaluates the effectiveness of using fuzzy sets to forecast daily Brent oil prices. It also contains possible modifications of the proposed method and in comparison with basic methods. The results suggest that Brent oil prices series have short memory because using information about last 2-days prices shows better forecast accuracy. Forecasting based on fixed universe of discourse shows better efficiency and it also proves that oil prices series has short memory. Adding the probability of switching between linguistic terms in defuzzification function could be used to improve accuracy of predictions. Also the approach can take into consideration expert's opinion about direction of future variation. The effective expert's work can reduce errors of forecast from 1.5% till 0.76%. But this modification can be used if experts correctly guess the direction of the change in trend in eight out of ten cases and more. The reasonable obtained results can be used by analysts dealing with the prediction of oil prices.

Keywords: petroleum prices, soft computing, Brent, fuzzy logic, time series, fuzzy sets, forecast, prediction, expert's opinion, stock market

1. Introduction

Oil is the most important energy resources in the world. Total turnover of oil takes a large share of world trade. Oil prices importantly concern to international organizations, governments, enterprises and different type of investors. Increase in oil prices on 10% influence on USA' GDP growth on 0,6-2,5% in research of Suater and Awerbuch (2006). Much has been said and written about influencing factors like weather, stock level, political games and psychological expectations. It has been establish by recent studies that high oil prices directly affect macroeconomic indicators such inflation in Cologni and Manera (2009), Farzanegan and Markwardt (2009), GDP in researches of Cologni and Manera, Doroodian and Boyd (2003), Prasad et al. (Prasad, Narayan, & Narayan, 2007), investments in researches of Abel (1990), Hamilton (2003), Rafiq et al. (Rafiq & Mallick, 2008) and generating financial crisis that confirmed in researches of Gisser and Goodwin (1986), Jones et al. (Jones, Leiby, & PAIK, 2004). These dependences have nonlinear and even chaotic character so that forecasting oil prices is difficult and always actual problem.

Contingently forecasting of oil prices can be divides into three types: short-term (1-3 days), middle-term (up to one year) and long-term forecasting (more than one year). Long-term forecasting usually bases on traditional econometric techniques such regression models that considering different factors as geopolitics, scientific and technical process, level of world oil reserves, social environment and other factors impacted on oil prices. Short-term and middle-term forecasting uses nonlinear and nonstationary time series models.

Carry out some overview about most popular techniques on short-term forecasting oil prices. Sadorsky (1999) uses several different univariate and multivariate statistical models to forecast oil volatility and show that forecast accuracy not associated with complexity of model. Models like state space, vector autoregression and bivariate GARCH do not perform as well as the single equation GARCH model.

Alexandra Costello et al. (Costello, Asem, & Gardner, 2008) compare the ARMA with historical simulation to the semi-parametric GARCH model. The results suggest that the semi-parametric GARCH model generates VaR forecasts that are superior to the VaR forecasts from the ARMA with historical simulation. This is due to the fact that GARCH captures volatility clustering. Weakness in GARCH model is based on using the normal assumption on the future risk dispersion. Marimoutou et al. propose Extreme Value Theory for forecasting VaR of Brent and WTI oil prices. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.