utilitiesCat: ~~ Internal utility functions and a user-visible constant for...

In FrF2: Fractional Factorial Designs with 2-Level Factors

~~ Internal utility functions and a user-visible constant for working with regular fractional factorials ~~

Description

~~ Internal functions, mainly for deriving words, alias structures and so forth from catalogue information for regular 2-level fractional factorials. They need not be accessed for normal use of the package - some of them may be made user-usable for future versions of the package.

The constant Yates is user-visible, as it can be useful for expert users. ~~

Usage

slowfast(k) 
mult.gen(liste) 
mult.gen.a(string.vec,sep="")
words.all(k, gen, design = NULL, max.length=7, select.catlg = catlg)
alias3fi(k, gen, order = 3)
gen.check(k, gen)
estimable.check(estimable, nfac, factor.names)
estimable(estimable, nfac, nruns, clear=FALSE, res3=FALSE, max.time=60, 
            select.catlg=catlg, perm.start=1:nfac, perms=NULL, order = 3, 
            method="VF2", sort="natural", ignore.dom=FALSE)
mapcalc(estimable, nfac, nruns, res3=FALSE, select.catlg=catlg, 
            method="VF2", sort="natural", ignore.dom=FALSE)
mapcalc.block(estimable, nfac, clearcur, 
            method="VF2", sort="natural")
mapcalc.distinct(estimable, nfac, nruns, res3=FALSE, max.time=60, 
        select.catlg=catlg, perm.start=1:nfac, perms=NULL)
map2design(map, select.catlg=catlg)
gen2clear.2fis(k, gen)
formula2matrix(formula)
char2num(char)
check.subisomorphic.special(estimable, nfac, hilf2, hilf3=NULL, 
    res3=FALSE, max.time=60, 
    perm.start=1:nfac, begin_time=Sys.time(), name=NA)
check.subisomorphic.matrix(estimable, nfac, hilf2, hilf3=NULL, 
    res3=FALSE, max.time=60, 
    begin_time=Sys.time(), name=NA, perms=perms)
getNext(perm)
Yates

Arguments

k	number of base factors spanning a full factorial with the desired number of runs (i.e. log2(nruns))
liste	list of words in terms of vectors of factor numbers to be combined into one new word
string.vec	vector of words in terms of factor letter combinations (e.g. “ABC”) to be combined into one new word
sep	character by which interaction terms are separated or concatenated; default is empty, i.e. ABC are separated into A, B and C; for more than 50~factors, a colon is needed so that F1:F2:F127 can be separated into F1, F2 and F127, for example.
gen	generators in any admissible format (gen.check transforms any admissible format to the default needed for calculations))
design	character string that contains a name of a design from select.catlg, like in FrF2
max.length	the maximum word length to be calculated for the words list; in large design, choosing this to be 4 or 5 may substantially speed things up!
select.catlg	catalogue of fractional factorial 2-level designs of class catlg
order	2 or 3, order of interactions to be shown
estimable	like in FrF2
nfac	number of factors, like nfactors in FrF2
factor.names	a named list with nfac elements, for which the names represent factor names
nruns	like in FrF2
clear	like in FrF2
res3	like in FrF2
max.time	like in FrF2
perm.start	like in FrF2
perms	like in FrF2
method	like in FrF2
sort	like in FrF2
ignore.dom	per default (ignore.dom=FALSE), non-dominating elements of select.catlg will be removed before searching; set ignore.dom to TRUE to change that (needed for combination of blocking with estimable requirement)
clearcur	is the set of clear 2fis after removing aliases with the block factor
map	map of experiment factors to design factors, used in estimable
formula	formula to be translated into matrix for estimable
char	character vector to be translated into matrix for estimable
hilf2	list of length four words
hilf3	list of length three words, if there are any
begin_time	timing option used for controlling overall run time, calculated
name	name of design in catalogue, used in case of max.time end for information where the stop occurred
perm	current permutation, of which getNext determines the next permutation in lexicographic order

Details

Yates is a user-visible constant that is useful in design construction:

Yates is a list of design column generators in Yates order (for 4096 runs), e.g. Yates[1:8] is identical to

list(1,2,c(1,2),3,c(1,3),c(2,3),c(1,2,3)).

Value

mult.gen and mult.gen.a determine a new word as the product of the argument words (numeric vector or character string respectively), words.all calculates a word list object with word length pattern, the word list itself and a max.length entry, alias3fi calculates an alias pattern based on the structure of the design (less demanding than the word list approach.

gen.check produces a list of vectors with integers from 1 to k for denoting the generators, which may have come as a vector of factor letter combinations or a vector of column numbers in Yates order, estimable.check determines a list including elements estimable, a 2-row-matrix indicating estimable 2fis from the permissible inputs, and nfac, the number of factors from input parameter estimable, function estimable calculates the appropriate design if certain effects have been requested to be estimable, and all subsequent functions are auxiliary functions to function estimable.

Yates is a user-visible constant (cf. details section).

Note

This package is still under development, but does already provide useful and well-tested results.

Author(s)

Ulrike Groemping

References

Hedayat, A.S., Sloane, N.J.A. and Stufken, J. (1999) Orthogonal Arrays: Theory and Applications, Springer, New York.

See Also

See also FrF2

FrF2 documentation built on Sept. 20, 2023, 9:08 a.m.