Forecasting Crude Oil Demand Using a Hybrid SVR and Markov Approach

Forecasting Crude Oil Demand Using a Hybrid SVR and Markov Approach

Wei Xu (Chinese Academy of Sciences, China), Jue Wang (Chinese Academy of Sciences, China) and Jian Ma (City University of Hong Kong, Hong Kong)
DOI: 10.4018/978-1-61520-629-2.ch012
OnDemand PDF Download:
No Current Special Offers


In this chapter, a hybrid support vector regression (SVR) and Markov forecasting approach is proposed for energy demand prediction. The original time series of energy demand is firstly decomposed into the general trend series and the random fluctuation series. Then the SVR method is used to model the general trend series and the Markov forecasting method is used to model the random fluctuation series so that the tendencies of two series can be accurately predicted. The prediction results of two series are integrated to formulate an ensemble output for future energy demand. The proposed forecasting approach makes full use of the historical time series information so as to improve the forecasting precision of time series with large random fluctuation. To illustrate the applicability and capability of the proposed approach, it is used to analyze and forecast world crude oil demand. For verification, the proposed approach is compared with SVR method, Markov forecasting method and ARIMA. The results show that the hybrid SVR and Markov forecasting approach proposed in this chapter can be applied successfully and provide high accuracy and reliability for forecasting world crude oil demand.
Chapter Preview


Crude oil plays an increasingly significant role in world economic development and also in the generation of wealth since nearly two-thirds of the world’s energy consumption comes from crude oil and natural gas. Worldwide consumption of crude oil exceeds $500 billion—roughly 10% of US GDP and crude oil is also the world’s largest and most actively traded commodity, accounting for about 10% of total world trade (Yu et al. 2008). World crude oil demand increased from 46 million barrels per day in 1970 to 85 million barrels per day in 2007. The growth rate of world crude oil demand from 1970 to 2002 averaged about 1.6% per year and world oil market was shocked by the strong demand growth that began in 2003 and continued through 2004 into 2007 (see Figure 1).

Figure 1.

The growth rate of world crude oil demand


Numerous factors (e.g. the growth of global economy, the growth of world crude oil stock and increasing load of coal and other energy) affect the growth of world crude oil demand, and the salience of each individual factor varies from time to time, so the time series of world crude oil demand shows large random fluctuation. How to analyze the fluctuation of world crude oil demand, and effectually to forecast world crude oil demand and supply in future have attracted many scholars and practicers’ attention.

A number of forecasting methods have been developed to forecast energy demand by many experts. Traditionally, econometric methods such as VAR models were developed to energy demand prediction. For example, Mackay et al. (1995) proposed a modified logit model to predict crude oil and natural gas. Rao et al. (1996) offered an econometric model to forecast petroleum demand. Crompton et al. (2005) suggested a BVAR method to analyze the trend of energy demand, and Ghosh (2006) provided a VECM method. However, Ediger et al. (2007) indicated that the estimated parameters in econometric models usually deviated from the realizations, and hence, time series forecasting appear to give better results and time series forecasting approaches such as ARIMA models were offered to forecast energy demand. For example, Chavez et al. (1999) suggested an ARIMA model to forecast energy consumption, and Ediger et al. (2007) proposed an ARIMA model to forecast primary energy demand. As Mentioned in Refenes et al. (1994), traditional forecasting techniques have reached their limitation in practical applications with nonlinearities in the dataset such as stock indices. Similarity, for the volatile energy demand, traditional statistical modeling is also insufficient since it is hard to capture the nonlinearity hidden in the energy demand.

As a result, many emerging artificial intelligence techniques, such as artificial neural networks (ANN), were used in energy demand forecasting and obtained good prediction performance. For example, Canyurt et al. (2004) suggested a GA approach to forecast energy demand. Murat et al. (2006) forecasted Transport energy demand using an ANN approach. In this paper, a hybrid forecasting approach based on SVR and Markov methods is proposed for crude oil demand prediction.

Complete Chapter List

Search this Book: