An Evaluation of Errors in Energy Forecasts by the SARFIMA Model
Leila Sakhabakhsh

Abstract
Forecasting is tricky business. This is particularly true in the energy field, where the highly random behavior of energy prices and technological change make forecasting difficult. In this paper, we study the long memory characteristic using Higuchi method and establish SARFIMA model to forecast consumption of petroleum products in U.S. Furthermore, we apply an error decomposition technique to study errors in energy forecasts by the SARFIMA model. The results indicate the SARFIMA model give precise predictions.

Full Text: PDF