An Autoregressive Integrated Moving Average Models For Process Output And Forecasting

In the first stage of analysis, some descriptive measures were used to examine the main properties of the series. Twelve-month centred moving average and differencing of order twelve advocated by Box and Jenkins (1970) were used to reduce the seasonal effect and to make the series stationary respectively. The autocorrelation of the differenced series of the productions were calculated including the partial autocorrelation with their respective correlogram which gives an insight into the probability model that generated the series.The second stage involves fitting of model and diagnostic checking to test for the adequacy of the fitted model. The diagnostic checking revealed that autoregressive model of order two was fitted into the series.