convert.to.rsm(fac.df, factors, numeric.levels, n.cp = min(n.factors + 1,4), alpha = nrow(fac.df)^(1/4), logx = F, keep.all = TRUE)
fac.design
frame, or a design with only two level factors.
fac.df
. These factors will be
used to construct the
rsm.design
. By default, all two level factors are used.
factors
. If this is omitted, default values for the two levels are 1
and 2.
rsm.design
for details.
TRUE
if all variables in
fac.df
are in the
returned
rsm.design
;
FALSE
if only the variables given by
factor
.
rsm.design
object derived from
fac.df
, with additional rows for
star and center points. The returned object is a central composite
design. The factors of
fac.df
are converted to
rsm.factor
s, that is, continuous variables with center points and
star points added. If
keep.all=TRUE
the non-factor variables are
augmented with NAs.
This function helps when an initial screening experiment is to be augmented with additional points to allow fitting a quadratic response surface.
bufferr.df <- convert.to.rsm(buffer.df, numeric = c(-1,1), keep = F) temp1 <- rsm.design(5, names(buffer.df)[1:5], fraction = 1/2) all.equal(bufferr.df, temp1) convert.to.rsm(design.digest('ff0308'), numeric = list(A = c(-1,1), B = c(10,20), C = c(0,1)))