At the right hand side of the Model Viewer window is a vertical toolbar to select the viewâ€”that is, the kind of plot or table that the viewer displays. Corresponding menu items appear under View in the menu bar. The function of each icon is explained in the following list.
Model Predictions
displays a plot of actual series values and model predictions over time. Click individual points in the graph to get a display
of the type (actual or predicted), ID value, and data value in the upper right corner of the window.
Prediction Errors
displays a plot of model prediction errors (residuals) over time. Click individual points in the graph to get a display of
the prediction error value in the upper right corner of the window.
Prediction Error Autocorrelations
displays horizontal bar charts of the sample autocorrelation, partial autocorrelation, and inverse autocorrelation functions
for the model prediction errors. Overlaid line plots represent confidence limits computed at plus and minus two standard errors.
Click any of the bars to display its value.
Prediction Error Tests
displays horizontal bar charts that represent results of white noise and stationarity tests on the model prediction errors.
The first bar chart shows the significance probability of the Ljung-Box chi-square statistic computed on autocorrelations
up to the given lag. Longer bars favor rejection of the null hypothesis that the series is white noise. Click any of the bars
to display an interpretation.
The second bar chart shows tests of stationarity of the model prediction errors, where longer bars favor the conclusion that the series is stationary. Each bar displays the significance probability of the augmented Dickey-Fuller unit root test to the given autoregressive lag. Long bars represent higher levels of significance against the null hypothesis that the series contains a unit root. For seasonal data, a third bar chart appears for seasonal root tests. Click on any of the bars to display an interpretation.
Parameter Estimates
displays a table showing model parameter estimates along with standard errors and t tests for the null hypothesis that the parameter is zero.
Statistics of Fit
displays a table of statistics of fit for the selected model. The set of statistics shown can be changed by using the Statistics
of Fit item under Options in the menu bar.
Forecast Graph
displays a plot of actual and predicted values for the series data range, followed by a horizontal reference line and forecasted
values with confidence limits. Click individual points in the graph to get a display of the type, date/time, and value of
the data point in the upper right corner of the window.
Forecast Table
displays a data table with columns containing the date/time, actual, predicted, error (residual), lower confidence limit,
and upper confidence limit values, together with any predictor series.