Current adjustment (AO approach)

The Current adjustment (AO approach) strategy means that the ARIMA model, outliers and other regression variables are not re-identified and the values of the associated coefficients are fixed. The transformation type also remains unchanged.

All new observations are automatically classified as additive outliers and corresponding coefficients are estimated during the regression phase. This does not impact the other variables’ coefficient estimation (they are fixed). The new observations are processed as intervention variables and the AO statuses will last till the next concurrent refreshment.

The example below shows the initial model (left) and the results of the refreshment procedure with the Current adjustment (AO approach) option (right). No element of the previous model has changed except all coefficient values are now fixed, hence the absence of p-values, scores and of the correlation of estimates matrix in the Summary. The original outlier (April 2006) is still in the refreshed model, with its fixed coefficient (its status went from “pre-specified” to “fixed”). The only difference is the addition of an intervention variable for the added month, September 2020. It can be seen at the end of the Summary and in the Regressor panel.

The comparison of the results from the initial nad refreshed models

Display-wise, there is an additional change. For the initial model the regressiors are displayed in the Regression panel.

The Regressor panel for the initial model

For the model refreshed with the the Current adjustment (AO approach) the Regressor panel does not contain any of the previous regressors (because they are fixed). However, it is easy to verify via the Summary that they are still in the model.

The Regressor panel for the refreshed model