monte {ssExtra} | R Documentation |
monteDoubleSampling
"
This is the ‘phantom’ generic for generating
objects of class "monteDoubleSampling
." Phantom because in
reality the generic is defined in sampSurf and this
documentation just links that generic to the double sampling
framework; i.e., there is no new monte
generic definition in
package ssExtra. There are currently two methods corresponding
to this generic for double sampling, the main one being for
objects of class ssBigBAF
; the full documentation
for both methods may be found in monte-methods
.
monte(object, ...)
object |
Signature argument, which differs for each method. This specifies the population from which samples will be drawn. |
... |
See methods. |
The methods associated with this generic should be used to
construct objects of class "monteDoubleSampling
."
These objects are specifically designed to hold information about
Monte Carlo experiments where one resamples from a known population to
infer efficiency and perhaps locate any bias in different sampling
estimators. The constructor methods can be used to look at traditional
normal theory and bootstrap confidence intervals in terms of nominal
capture rates for the population mean.
As already noted, the monte
generic is actually defined in
sampSurf (see monte
for
details) and has been appropriated here for use in double sampling
designs. The structure of the double sampling objects is quite
different than those for the original sampSurf version. Thus,
there is no new monte
generic defined, but there are new
methods defined for the generic in this package for double sampling
designs. See the returned value below for the difference.
A valid object of class “monteDoubleSampling
”.
Please note that there are two methods currently associated with the
monte
generic that can be used for double sampling. However,
only the main method whose signature argument is of class
“ssBigBAF” should normally be called by the user. The second
constructor method takes and object of class
“monteBigBAF”. This may seem a little strange, as this is the
class that we are actually building in the first constructor. The key
here is that the point-based Delta method (PBDM), or
“Lynch's method”, for big BAF sampling variance
estimation, was not developed until after the main code was
written. It seemed best not to add to the rather long original
constructor method for “ssBigBAF” objects or the code would be
very difficult to follow, indeed much more so than it already
is. Also, adding the code through a simple function would have
entailed passing a large number of arguments, or a a large list
object to the function to add the required components for the
PBDM. The simplest strategy was to build an almost
complete object with everything but the PBDM variance
components in the main (“ssBigBAF”-based) constructor, pass
this invalid object to the “monteBigBAF”-based
constructor method, and return from that a valid object of
class “monteBigBAF”.
This process does indeed work well. Please note that calling the
“monteBigBAF”-based constructor with a valid object will not
hurt anything as it will simply write over the contents of the
PBDM additions. So if one is inclined to do so, one can with
no harm done other than a little wasted CPU time; and there
may be some occasion to do so, but it is largely discouraged. The
upshot is that one should simply use the monte
constructor
method with signature for “ssBigBAF” objects and that is all
that is necessary to construct a valid object of
class“monteBigBAF”.
Jeffrey H. Gove
The ‘“monte”: When is n Sufficiently Large?’
vignette in sampSurf.
Please also see the User's Guide vignette for the ssExtra package.
monte
,
monteDoubleSampling
,
monte-methods
# # this example is commented-out because it consumes a lot of # cpu time when checking the package--just copy and paste # it if you want to try them out... # # from an existing ssBigBAF object wpAll520.bb that has # surfaces with count BAF = 5, and big BAF = 20... # ## Not run: wpAll520.mc = monte(wpAll520.bb, n=c(10,25,50), start=355, mcS=50) hist(wpAll520.mc) ## End(Not run)