Document Type : Research Paper

Authors

• masoud fazlalipoour ¹
• parviz nasiri ²

¹ Academic faculty member

² Department of Statistics, Payam Noor University, P.O. 19395-4697, Tehran, Iran

Abstract

It is of particular importance to examine statistical models for fitting time series data and provide a suitable model and predict important elements in macroeconomic and financial planning. One of these time series models that is widely used in the analysis of economic, meteorological, geographical and financial data is the ARFIMA model. In this model and other time series models, the parameters of the model are estimated by assuming that the error term has a normal distribution. In this article, while examining the behavior of the ARFIMA model, the Bayesian estimation of the fractional difference parameter is estimated by considering the appropriate prior distribution for it. Then, using oil export data, the deficit difference parameter of the ARFIMA model is estimated by considering the appropriate prior distribution. Finally, the goodness of fit of the ARFIMA model is compared with the models that are introduced through the RMSE criteria and it is shown that the model estimated by the Bayesian method has a better performance.

Keywords

• Time series data
• evaluation criteria
• fractional interval parameter
• Bayesian estimation
• prediction model

Main Subjects

• مطالعات اقتصادی مرتبط با حامل‌های انرژی (فسیلی، تجدیدپذیر و برق)