Author(s): TI Ayakeme, OE Biu, D Enegesele and N Wonu

Abstract: The study examined time series models for forecast comparison between ARIMA model and Winter methods (Additive and Multiplicative methods) of Bayelsa State Government Internally Generated Revenue (IGR) data frame of 2012 to 2018. The data used for the study were analyzed using Statistical software packages: MINITAB 18 and Micro software Excel. For a better understanding of the behavior of the data, monthly and yearly plots were done. The plot shows seasonality of order 4 with a regular first difference. Several ARMA (p, q) Models were fitted to the internally generated revenue with respective residuals as having white noise. The identified ARIMA model was ARIMA (0, 1, 1) (0, 0, 0)₄. (X_t =1.1524X_t-1 -0.1524X_t-2 + 0.7846e_t-1 +e_e) which has the least values of AIC and BIC amongst the fitted models. The identified ARIMA model was used to predict for 2019 to 2021. Furthermore, winters (additive and multiplicative) methods were employed in the study to model and forecast the internally generated revenue. The winters method models forecast values were compared with the identified ARIMA model forecasts using the Mean Absolute Error (MAE) and Mean square error (MSE). The ARIMA model forecast is better than winter methods: additive and multiplicative methods and was thus recommended for use.

Pages: 107-116 | Views: 670 | Downloads: 16

Download Full Article: Click Here