Forecasting Crude Oil Price with an Autoregressive Integrated Moving Average (ARIMA) Model
In this chapter, an autoregressive integrated moving average (ARIMA) model is proposed to predict world crude oil. Data from 1970 to 2006 is used for model development. We find that the model is able to describe and predict the average annual price of wor
- PDF / 976,740 Bytes
- 12 Pages / 439.37 x 666.142 pts Page_size
- 49 Downloads / 303 Views
Abstract In this chapter, an autoregressive integrated moving average (ARIMA) model is proposed to predict world crude oil. Data from 1970 to 2006 is used for model development. We find that the model is able to describe and predict the average annual price of world crude oil with the aid of SAS software. The mean absolute percentage error (MAPE) is 4.059 %. Experiment shows the model have the preferable approach ability and predication performance, particularly for the short - term forecast. Keywords Crude oil price forecast · ARIMA model · SAS software.
1 Introduction Crude oil has been playing an increasingly important role in the world economy since nearly two-third of the world’s energy demands is met from crude oil [1]. For example, central banks and private sector forecasters view the price of oil as one of the key variables in generating macroeconomic projections and in assessing macroeconomic risks. Forecast of the price of oil, play a role in generating projections of energy use, in modeling investment decisions in the energy sector, in predicting carbon emissions and climate change, and in designing regulatory policies such as automotive fuel standards or gasoline taxes [2]. However, crude oil price forecasting is a very important topic, albeit an extremely hard one, due to its intrinsic difficulties and high volatility [3]. The average annual price of world crude oil series can be considered as C. Zhao (B) · B. Wang School of Science, Southwest Petroleum University, Chengdu 610500, China e-mail: [email protected] B. Wang e-mail: [email protected]
B.-Y. Cao and H. Nasseri (eds.), Fuzzy Information & Engineering and Operations Research & Management, Advances in Intelligent Systems and Computing 211, DOI: 10.1007/978-3-642-38667-1_27, © Springer-Verlag Berlin Heidelberg 2014
275
276
C. Zhao and B. Wang
a nonlinear and non-stationary time series, which is interactively affected by many factors, predicting it accurately is rather challenging. In the past decades, traditional statistical and econometric techniques, such as linear regression (LinR), co-integration analysis, GARCH models, naive random walk, vector auto-regression (VAR) and error correction models (ECM) have been widely applied to crude oil price forecasting [1]. In 1994, Huntington applied a sophisticated econometric model to predict crude oil prices in the 1980s [4]. In 1995, Abramson and Finizza utilized a probabilistic model for predicting oil prices [5]. In 2001, Morana suggested a semiparametric statistical method for short-term oil price forecasting based on the GARCH properties of crude oil price [6]. Similarly, in 1998, Barone-Adesi et al. suggested a semiparametric approach for oil price forecasting [7]. In 1988, Gulen used co-integration analysis to predict the West Texas Intermediate (WTI) price [8]. In 2002, 2005 and 2006, Ye, M. et al presented a simple econometric model of WTI prices, using OECD petroleuminventory levels, relative inventories, and high-inventory and lowinventory variables [9–11]. In 2004, Mirmirani and
Data Loading...
