predict.DSFM2D: Predict Method for Two-Dimensional DSFM Fits

In MarcGumowski/dysefamor: Estimation Algorithm of Dynamic Semiparametric Factor Model

Description

Predicted values based on "DSFM2D" object using Vector Autoregressive Processes (VAR(p)).

Usage

## S3 method for class 'DSFM2D'
predict(object, nAhead = 12,
  x1Forecast = unique(object$mHat[, 1]), x2Forecast = unique(object$mHat[,
  2]), p = 1, ...)

Arguments

object	an object of class "DSFM1D".
nAhead	the number of steps ahead for which prediction is required.
x1Forecast	the vector of covariates in the first dimension to be forecasted.
x2Forecast	the vector of covariates in the second dimension to be forecasted.
p	the order of the Vector Autoregressive Process to be fitted.
...	other parameters to be passed through the VAR and predict.varest functions.

Details

This function makes uses of package vars to fit and predict VAR(p) processes, before pluging in the predicted factor loadings in the DSFM. The factors functions are interpolated by the function interp of package akima.

Value

predict.DSFM2D returns an object of class "predict.DSFM2D". This class is a list containing:

YHatForecast	the forecasted responses.
YHatForecastList	the forecasted responses in a more usual format.
ZHatForecast	the predicted factors loadings.
nAhead	the number of steps ahead.

References

Bernhard Pfaff (2008). VAR, SVAR and SVEC Models: Implementation Within R Package vars. In: Journal of Statistical Software 27(4). URL http://www.jstatsoft.org/v27/i04/.

Hiroshi Akima and Albrecht Gebhardt (2015). akima: Interpolation of Irregularly and Regularly Spaced Data. R package version 0.5-12. URL http://CRAN.R-project.org/package=akima

See Also

VAR,predict.varest, interp, DSFM, DSFM2D, DSFM2DData.

MarcGumowski/dysefamor documentation built on May 7, 2019, 2:47 p.m.