This function allows to inspect the list of available orthogonal arrays, optionally specifying selection criteria
show.oas(name = "all", nruns = "all", nlevels = "all", factors = "all",
regular = "all", GRgt3 = c("all", "tot", "ind"), Rgt3 = FALSE, show = 10,
parents.only = FALSE, showGRs = FALSE, showmetrics = FALSE, digits = 3)
character string or vector of character strings giving name(s) of (an) orthogonal array(s); results in an error if name does not contain any valid name; warns if name contains any invalid name
the requested number of runs or a 2-element vector with a minimum and maximum for the number of runs
a vector of requested numbers of levels for a set of factors in question, must contain integers > 1 only; nlevels cannot be specified together with factors
a list with the two elements nlevels
and number
, which are both integer
vectors of equal length;
nlevels
contains the number of levels and number
the number of factors
for the corresponding number of levels
either unrestricted (the default “all”), a logical which requests
(TRUE
) or rejects (FALSE
) regular arrays, or the character string
“strict” to request strictly regular arrays, for which all confounded factors
are completely confounded with a 2-factor interaction of two other factors (the
latter are fixed level arrays or crossed arrays)
either unrestricted (the default “all”), or a character string which requests
GR
("tot"
) or GRind
("ind"
) to be larger than 3
an integer number specifying how many arrays are to be listed (upper bound),
or the character string "all"
for showing all arrays, no matter how many.
The default is to show 10 arrays. show = 0
switches off the display of the result
and only returns a value. Since August 2018, the number refers to stronger and weaker arrays, separately.
logical specifying whether to show only parent arrays or child arrays as well;
the default is FALSE
for inclusion of child arrays
logical specifying whether to show the generalized resolution quality metrics
with the resulting arrays; the default is FALSE
. If set to TRUE
,
three metrics are displayed (see Details section).
logical specifying whether to show all array quality metrics with the resulting
arrays; the default is FALSE
. If set to TRUE
, several metrics
are displayed (see Details section).
integer number of significant digits to show for GR and A metrics;
irrelevant, if showmetrics
is FALSE
A data frame with the three columns name
, nruns
and lineage
,
containing the array name, the number of runs and - if applicable - the lineage for generating the array
from other arrays. The lineage entry is empty for parent arrays that are either directly available
in the package and can be accessed by giving their name (e.g. L18.3.6.6.1
) or are full factorials
(e.g. L28.4.1.7.1
). If further information has been requested (e.g. with showmetrics=TRUE
),
the data frame contains additional columns.
If no array has been found, the returned value is NULL
.
The function shows the arrays that are listed in the data frames oacat
or oacat3
.
For child arrays that have to be generated with a lineage rule
(can be automatically done with function oa.design
), the lineage is displayed
together with the array name. The option parent.only = TRUE
suppresses printing and output of child arrays. The structure of the lineage entry
is documented under oacat
.
If display of metrics is requested with showmetrics=TRUE
, the printed output shows the metrics
GR*, GRind*, regular (logical, whether regular or not), SCones* (number of squared canonical correlations that are 1),
and the numbers of words of lengths 3 to 8 (A3 to A8). showGRs=TRUE
requests the metrics marked with asterisks only (without SCones
in case GRgt3="ind"
). More information on
all these metrics can be found here
Kuhfeld, W. (2009). Orthogonal arrays. Website courtesy of SAS Institute https://support.sas.com/techsup/technote/ts723b.pdf and references therein.
Mee, R. (2009). A Comprehensive Guide to Factorial Two-Level Experimentation. New York: Springer.
oa.design
for using the arrays from oacat
in design creation
oacat
for the data frames underlying the function
# NOT RUN {
## the first 10 orthogonal arrays with 24 to 28 runs
show.oas(nruns = c(24,28))
## the first 10 orthogonal arrays with 24 to 28 runs
## excluding child arrays
show.oas(nruns = c(24,28), parents.only=TRUE)
## the orthogonal arrays with 4 2-level factors, one 4-level factor and one 5-level factor
show.oas(factors = list(nlevels=c(2,4,5),number=c(4,1,1)))
## show them all with quality metrics
show.oas(factors = list(nlevels=c(2,4,5),number=c(4,1,1)), show=Inf, showmetrics=TRUE)
## pick only those with no complete confounding of any degrees of freedom
show.oas(factors = list(nlevels=c(2,4,5),number=c(4,1,1)), GRgt3="ind", showmetrics=TRUE)
## the orthogonal arrays with 4 2-level factors, one 7-level factor and one 5-level factor
show.oas(factors = list(nlevels=c(2,7,5),number=c(4,1,1)))
## the latter orthogonal arrays with the nlevels notation
## (that can also be used in a call to oa.design subsequently)
show.oas(nlevels = c(2,7,2,2,5,2))
## calling arrays by name
show.oas(name=c("L12.2.11", "L18.2.1.3.7"))
# }
Run the code above in your browser using DataLab