The tool supports interface of Foresight Analytics Platform 9 or earlier.
The panel is used to edit basic model parameters:
Type of Error Correction Model. The error correction model is determined:
Without trend in autoregression, without constant in cointegration equation.
Without trend in autoregression, with constant in cointegration equation.
With constant in autoregression and in cointegration equation (trend in source data).
With trend in autoregression, with constant and trend in cointegration equation.
With quadratic trend in autoregression, with constant and trend in cointegration equation.
Number of Cointegration Relations. The number of cointegration relations in the model is determined. This value should be less than the number of endogenous variables but greater or equal to one. The exact number of cointegration equations is usually not known beforehand and must be estimated, for example, using the Johansen test.
Endogenous Variables. Endogenous variables can be added, deleted or transformed in the same way as input variables (factors). A set of exogenous variables is determined for each endogenous variable. When a variable is selected, a list of its exogenous variables is displayed below.
Autoregression Order. Set the autoregression order for endogenous variables in the box. Enter numbers or ranges of autoregression order separated by commas. The range of autoregression order is specified using the "-" character.
Exogenous Variables. Exogenous variables can be added, deleted or transformed, and their lag can be set in the same way as input variables (factors).
NOTE. Variables are divided into endogenous and exogenous depending on the model's concept. it is assumed that in each equation, exogenous variables should not be correlated with error while endogenous variables should be correlated with error.
Forecast Correction. Selecting the checkbox enables the user to set the variable used for forecast correction. The variable is not included into the identified equation of the model. After the checkbox is selected, the Create button is enabled, the button creates a variable (without data) that is used to adjust forecast. The variable frequency corresponds to model frequency. The created variable is located in the root folder of the modeling container. It is named Forecast Correction Factor and automatically opens for edit.
When the variable is selected, which number of dimensions does not match the number of dimensions of the output variable, the Change Dimension dialog box opens, which is used to fix the dimensions missing in the output variable. The dialog box can be opened again by clicking the Fix button.
NOTE. If a time series database is the parent object of the modeling container, variables on attributes are used for model calculation. Working with variables is described in the Model Parameters section. A variable cannot be created or fixed.
See also: