Sets suitable defaults for the spotConfig part of TDMR.
defaultSC()
a list with the following elements (the values in parantheses []
are the defaults):
["NEEDS_TO_BE_SET"] a data frame with columns lower
,
upper
, type
, row.names
, each a vector with as many
entries as there are parameter to be tuned
["NEEDS_TO_BE_SET"]
["NEEDS_TO_BE_SET.conf"] a string ending on ".conf", the configuration name
[FALSE] see spotControl
[FALSE] TRUE: make a line plot showing progress
[1] see spotControl
[50] the budget, max number of algo evaluations
[designLHD] function that creates initial design,
see spotControl
[10] number of initial design points (former init.design.size)
[2] number of initial repeats (former init.design.repeats)
[2] number of repeats for the same model design point
[TRUE] whether the object function has noise or not
(Note: TRUE is required if replicates>1
(!))
[mean] how to merge Y over replicates: mean or min
[buildKriging] function that builds the surrogate model,
see spotControl
[optimLHD] function that optimizes on surrrogate,
see spotControl
[100] optimizer budget (former seq.design.size)
[2] optimLHD retries (former seq.design.retries)
With the call setParams(mySC,defaultSC())
it is possible to extend a partial list
mySC
to a list containing all sC
-elements (the missing ones are taken from
defaultSC()
). If mySC
has an element not present in defaultSC()
,
this element is not taken and a warning is issued.
With setParams(mySC,defaultSC(),keepNotMatching=TRUE)
also elements
of mySC
not present in defaultSC()
are taken (no warnings).