Application of Unit Root Tests in Seasonal Time Series Prediction"The Case of the Retail Price of Meat Commodity Groups in Iran"

Document Type : Research Paper

Authors

Abstract

In this paper, meanwhile introducing seasonal unit root test HEGY, we used this test for determining seasonal and non-seasonal unit roots of retail price time series for four  of meat products: chicken, salmon, shrimp and beef. For this purpose, the monthly data on retail prices of these products are used across the country for the years 1380-86. The results showed that all price series except chicken meat retail price besides the unit root at zero frequency, have unit root in one or several unit root in seasonal frequencies besides being the result of nonstationary random seasonal process in all series confirm that the prosper difference filter for their stationary are different from seasonal difference filters that was proposed in box and jenkinz approach. Based on HEGY test, traditional methods in experimental studies because use of all the roots of the seasonal (default seasonal roots at all frequencies), causing loss of series inner information and making stipulation bias. This study seeks the time series of monthly behavior in each seasonal frequency separately by HEGY test rather than the default placement of unit roots at all frequencies of occurrence will avoid illusory results.

Keywords