Author(s): David Peter Oaya, Nweze Obini Nwanze and Abubakar Muhammad Auwal

Abstract: The researched on the efficiency of establishing of variance obtained from Seasonal Autoregressive Integrated Moving Averages (SARIMA) model for resampled and actual inflation time series data. The study used the monthly inflation data of 2014 to 2018. It conducting stationarity test on the time series data using Augmented Dickey-Fuller (ADF) unit roots test in order to avoid wrong forecast. Fitting a model on resampled and non-resampled time series data was important to get accurate forecasts and predictions, as well as, check the precision between the variance of these models and outcomes. Using the Akaike Information Criterion (AIC), most efficient forecast was obtained. The Seasonal Autoregressive integrated moving averages (SARIMA) was applied to both the stationary data for bootstrap and non-bootstrap samples. Results show that the models fit on resampled inflation series produced higher AIC value of 689.05 of SARIMA (3,0,1) (3,1,0)12 compared to the actual inflation data having an AIC value of 269.9 of SARIMA (2,0,0) (2,0,0)12. The findings reveal that the SARIMA model of the actual inflation data was more perfect for forecasting compared to that of the resampled data. The inflation data became stationary at the sixth difference while that of resampled data became stationary at second difference. The study recommends that it is ideal and better to use actual data of seasonal time series event for forecasting.

Pages: 08-11 | Views: 449 | Downloads: 18

Download Full Article: Click Here