R: Compare Forecasts Cov Relative to True Model Output

forecastCov.estimatorsWRTtrue {dse2}

R Documentation

Compare Forecasts Cov Relative to True Model Output

Description

Usage

    forecastCov.estimatorsWRTtrue(true.model, Spawn=.SPAWN, rng=NULL,
                       simulation.args=NULL,
                       est.replications = 2, pred.replications = 2,
                       discard.before = 10, horizons = 1:12, quiet = FALSE,
                       estimation.methods=NULL, compiled=.DSECOMPILED)
    is.forecastCov.estimatorsWRTtrue(obj)

Arguments

`true.model`	An object of class TSmodel.
`estimation.methods`	A list as used by estimate.models.
`simulation.args`	an arguments to be passed to simulate.
`est.replications`	An arguments to be passed to simulate.
`pred.replications`	An arguments to be passed to simulate.
`discard.before`	An integer indicating the number of points in the beginning of forecasts to discard for calculating covariances.
`zero`	If TRUE then forecastCov is also calculated for a forecast of zero.
`trend`	If TRUE then forecastCov is also calculated for a forecast of a linear trend.
`horizons`	Horizons for which forecast covariance should be calculated.
`rng`	If specified then it is used to set RNG.
`Spawn`	If TRUE then Splus For loops are used.
`quiet`	If TRUE then some messages are not printed.
`compiled`	a logical indicating if the compiled version of the code should be used. (F would typically only be used for debugging.)
`obj`	an object.

Details

Calculate the forecasts cov of models estimated from simulations of true.model with estimation methods indicated by estimation.methods (see estimate.models). This function makes multiple calls to forecastCovWRTtrue.

Value

The returned results has element forecastCov.true, forecastCov.zero, forecastCov.trend containing covariances averaged over estimation replications and simulation replications (forecasts will not change but simulated data will). forecastCov a list of the same length as estimation.methods with each element containing covariances averaged over estimation replications and simulation replications. estimated.models a list of length est.replications, with each elements as returned by estimate.models, thus each element has multi.model as a subelement containing models for different estimation techniques. So, eg. estimated.models[[2]]$multi.model[[1]] in the result will be the model from the first estimation technique in the second replication.

Examples

if(is.R()) data("eg1.DSE.data.diff", package="dse1")
true.model <- est.VARX.ls(eg1.DSE.data.diff) # just to have a starting model
z <-  forecastCov.estimatorsWRTtrue(true.model, 
    estimation.methods=list(est.VARX.ls=list(max.lag=4)))

[Package Contents]