Object Reference : Object View and Procedure Reference : Equation
Examine ARMA structure of estimated equation.
Provides diagnostic graphical and tabular views that aid you in assessing the structure of the ARMA component of an estimated equation. The view is currently available only for equations specified by list and estimated by least squares that include at least one AR or MA term. There are four views types available: roots, correlogram, impulse response, and frequency spectrum.
eq_name.arma(type=arg [,options])
where eq_name is the name of an equation object specified by list, estimated by least squares, and contains at least one ARMA term.
Required “type=” option selects the type of ARMA structure output: “root” displays the inverse roots of the AR/MA characteristic polynomials, “acf” displays the second moments (autocorrelation and partial autocorrelation) for the data in the estimation sample and for the estimated model, “imp” displays the impulse responses., “freq” displays the frequency spectrum.
Displays the table view of the results for the view specified by the “type=” option. By default, EViews will display a graphical view of the ARMA results.
Specifies the maximum lag length for “type=acf”, and the maximum horizon (periods) for “type=imp”.
Specifies the size of the impulse for the impulse response (“type=imp”) view. By default, EViews will use the regression estimated standard error.
Stores the results as a matrix object with the specified name. The matrix holds the results roughly as displayed in the table view of the corresponding type. For “type=root”, roots for the AR and MA polynomials will be stored in separate matrices as NAME_AR and NAME_MA, where “NAME” is the name given by the “save=” option.
Force the dialog to appear from within a program.
Print the table or graph output.
eq1.arma(type=root, save=root)
displays and saves the ARMA roots from the estimated equation EQ1. The roots will be placed in the matrix object ROOT.
eq1.arma(type=acf, hrz=25, save=acf)
computes the second moments (autocorrelation and partial autocorrelations) for the observations in the sample and the estimated model. The results are computed for a 25 period horizon. We save the results in the matrix object ACF.
eq1.arma(type=imp, hrz=25, save=imp)
computes the 25 period impulse-response function implied by the estimated ARMA coefficients. EViews will use the default 1 standard error of the estimated equation as the shock, and will save the results in the matrix object IMP.
displays the frequency spectrum in graph form.
See “ARMA Structure” for details. See also “Time Series Regression”.