Utility functions for DoE packages, not intended for direct use
make.generators(name, liste)
make.formulas(orignames, factor.names)
des.recode(var, recodes, as.factor.result, char)
printBy(data, byvars, ...)
Letters
gen.fun(obj, num = FALSE)
generators(design, ...)
# S3 method for design
generators(design, ...)
# S3 method for catlg
generators(design, ...)
# S3 method for character
generators(design, select.catlg = catlg, ...)
PFTs.from.variants(array, variants, R=3, rela=TRUE)
matrix.fromPFTs(PFTs)
rankPFT(pfts)
bestPFT(pfts)
rect_simplex(r)
HouseholderRotToOne(from)Most functions are not intended for direct use.
The generators methods output a list of human-readable character string
vectors with the generators of catalogue entries or a design.
The following functions have not yet been adapted to also handle the more
attractive new metrics calculated with functions P3.3, P4.4 and
GRind.
Function PFTs.from.variants outputs a list of (relative) projection
frequency tables (each entry is produced with function P3.3 or P4.4).
Function matrix.fromPFTs outputs a matrix (reshaped from such a list, filling
missing entries with 0 counts for some designs, if required), with columns for
the different selections and rows for different (relative) numbers of generalized words.
Function rankPFTs outputs a ranking vector, with 1 for the best design
(ties are kept in the original order). Function bestPFTs outputs a logical vector,
with entry TRUE for all best designs (identical best (R)PFTs).
...
...
...
...
variable
character string specifying recodes, so far only used in the principal form “-1=level[1];1=level[2]”
logical that specifies whether or not results are to be coerced to factor. The default is that factors are coerced to factor, while numeric variables with numeric recodes are kept numeric.
logical: should replacement be treated as character string (TRUE) or evaluated? default is FALSE
data frame to be printed in groups that are defined by the byvars
character vector of variable names;
variables named here determine the groups; only adjacent observations are considered part of a group, i.e.
if data are not ordered by the byvars, several separate occurrences of the same group may occur
an entry of a class catlg object
logical; if TRUE, Yates matrix column numbers are output, otherwise human-friendly generators
FrF2 design given as catalogue entry, design itself or character string; function generators provides the generating contrasts for additional factors for these designs.
the catalogue in which to look for the design (should be of class catlg)
further arguments to function print or generators
orthogonal array (available in the workspace); must be matrix or data frame
matrix the rows of which contain column numbers for array
resolution of array and also array[variants,,drop=FALSE]
logical; relative PFTs (TRUE) or absolute PFTs (FALSE)
list of (absolute or relative) projection frequency tables.
The list could have been produced with function PFTs.from.variants.
a matrix the columns of which contain the PFT entries for several designs,
with row names indicating the (relative) number of words of a triple/quadruple.
The matrix could have been obtained from function matrix.fromPFTs.
integer; dimension of the simplex
numeric vector to be rotated towards (1,0,...,0)
Ulrike Groemping
Most functions are not intended for direct use.
The constant Letters is the vector of all capital and lower case
letters of the Roman alphabet except I and i.
It is used for default factor names and has been obtained as
Letters <- c(LETTERS[-9],letters[-9]).
The generic function generators and its methods determine generating contrasts
for regular fractional factorial 2-level designs, which can be specified as
the design itself (created with function FrF2),
a list of catalogue entries in package FrF2 object catlg,
or a character string that names such a catalogue entry.
The method for the design itself cannot be used on designs of type
FrF2.param... or FrF2.folded. Also, it cannot
be used on designs produced with versions of FrF2 before 1.1 in case
of blocked or splitplot designs. It is generally recommended to update all packages of the
suite (DoE.base, FrF2, DoE.wrapper, RcmdrPlugin.DoE), whenever
one of them is updated.
This paragraph describes internal functions: Function PFTs.from.variants
can be used to determine a list of (relative)
projection frequency tables (cf. Groemping 2011) for comparing several column
selections from the same array. Function matrix.fromPFTs brings
this list into matrix form, functions rankPFT and bestPFT
allow ranking or picking best selections.
Groemping, U. (2011). Relative projection frequency tables for orthogonal arrays. Report 1/2011, Reports in Mathematics, Physics and Chemistry http://www1.bht-berlin.de/FB_II/reports/welcome.htm, Department II, Berliner Hochschule fuer Technik (formerly Beuth University of Applied Sciences), Berlin.
Hedayat, A.S., Sloane, N.J.A. and Stufken, J. (1999) Orthogonal Arrays: Theory and Applications, Springer, New York.
See Also FrF2-package, DoE.wrapper-package
## default factor names for most design generating functions
## (some quantitative designs have other default factor names)
## for up to 50 factors
Letters
if (FALSE) {
generators("7-2.2")
generators(catlg[2:8])
generators(FrF2(16,7))
generators(FrF2(16,5,blocks=4,alias.block.2fi=TRUE))
generators(FrF2(16,5,WPs=4,nfac.WP=2))
}
## column selections from L18 with one 2-level and six 3-level factors
v <- rbind(1:7, c(1:6,8), c(1:5,7:8), c(1:4,6:8), c(1:3,5:8), c(1:2,4:8), c(1,3:8))
## RPFTs
RPFTs <- DoE.base:::PFTs.from.variants(L18, v, rela=TRUE)
rpfts <- DoE.base:::matrix.fromPFTs(RPFTs)
rpfts
DoE.base:::rankPFT(rpfts)
DoE.base:::bestPFT(rpfts)
Run the code above in your browser using DataLab