| distribution {dse2} | R Documentation |
Plot distribution of estimates.
distribution(obj, ...)
## S3 method for class 'TSdata':
distribution(obj, ..., bandwidth=0.2,
select.inputs = seq(length= nseriesInput(obj)),
select.outputs= seq(length=nseriesOutput(obj)))
## Default S3 method:
distribution(obj, ..., bandwidth=0.2, series=NULL)
## S3 method for class 'coefEstEval':
distribution(obj, ..., Sort=FALSE, bandwidth=0.2,
graphs.per.page=5)
## S3 method for class 'rootsEstEval':
distribution(obj, ..., mod=TRUE, invert=FALSE, Sort=FALSE,
bandwidth=0.2, select=NULL)
obj |
an object as returned by EstEval. |
Sort |
if Sort is true then sort is applied. This helps (a bit) with estimation methods like black.box which may not return parameters of the same length or in the same order. |
bandwidth |
passed to density or ksmooth. |
graphs.per.page |
integer indicating number of graphs to place on a page. |
series |
series to be plotted. (passed to selectSeries) |
select.inputs |
series to be plotted. (passed to selectSeries) |
select.outputs |
series to be plotted. (passed to selectSeries) |
... |
other objects to be plotted (not working for some methods). |
invert |
logical indicating if the inverse of roots should be plotted |
mod |
logical indicating if the modulus of roots should be plotted |
select |
integer vector indicating roots to be plotted. If select is not NULL then roots are sorted by magnitude and only the indicated roots are plotted. For example, select=c(1,2) will plot only the two largest roots. |
ksmooth is applied if available to get a smoothed
estimate of the distribution of the estimates. If ksmooth is not
available then density is applied if it is available.
None
data("eg1.DSE.data.diff", package="dse1")
model <- estVARXls(TSdata(output=outputData(eg1.DSE.data.diff)), max.lag=2)
# now use this as the true model
z <- EstEval(model,
estimation="estVARXls", estimation.args=list(max.lag=2))
distribution(z)
tfplot(z)