Page 69 - Azerbaijan State University of Economics

Murad Yusifov: Modelling the inflationary processes and forecasting:an application of ARIMA,
SARIMA models

series period. It is known that the most times series are non-stationary. It changes the integrated

order of the series. ARIMA model in generally can be written as follows.

(1)

ARIMA is non-seasonal times series model. The differencing linear operator in the model is

noted as .

(3)

The general form of the above mentioned ARIMA model with integrated order

can be defined as below mentioned:

(4)

(5)

(6)

ARIMA model is based on Box-Jenkins (BJ) methodology. This methodology consists of

four stages: Identification. Estimation. Diagnostic checking. Forecasting. In most cases such

models can generate the more reliable results. The figures , Akaike info criteria(AIC), SIC,

, , and obtained from the models should be compared and the model

having the least indicator should be considered the best model [8].

In this study seasonal adjusted ARIMA model (SARIMA) və non-seasonal ARIMA

model have been applied and the results are compared on diagnostic basis. Seasonal ARIMA

proseseses having the times series aşağıdakı

kimi ifadə olunur:

69