A package dedicated to the automatic generation of regular factorial designs, including fractional designs, orthogonal block designs, row-column designs and split-plots.

The user describes the factors to be controlled in the experiment and the anova model to be used when the results will be analysed. He or she also specifies the size of the design, that is, the number of available experimental units. Then planor looks for a design satisfying these specifications and possibly randomizes it. The core of the algorithm is the search for the key matrix, an integer matrix which determines the aliasing in the resulting factorial design.

The user may use the function `regular.design`

where all
these steps are integrated, and transparent by default. Alternatively,
the steps can be decomposed by using successively the functions
`planor.factors`

, `planor.model`

,
`planor.designkey`

and
`planor.design`

. For the expert
user, the function `planor.designkey`

can give several key
matrix solutions. Alias and summary methods allow to study and compare
these solutions, in order to select the most appropriate one for the
final design.

An R option named `planor.max.print`

is set. It is equal
to the number of printed
rows and columns in the display of planor matrices. Default is 20.
You can change its value by using the function
`options()`

(see `?options`

).

Monod, H. and Bouvier, A. and Kobilinsky, A.

See `citation("planor")`

.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | ```
# DESIGN SPECIFICATIONS
# Treatments: four 3-level factors A, B, C, D
# Units: 27 in 3 blocks of size 9
# Non-negligible factorial terms:
# block + A + B + C + D + A:B + A:C + A:D + B:C + B:D + C:D
# Factorial terms to estimate:
# A + B + C + D
# 1. DIRECT GENERATION, USING regular.design
mydesign <- regular.design(factors=c("block", LETTERS[1:4]),
nlevels=rep(3,5), model=~block+(A+B+C+D)^2, estimate=~A+B+C+D,
nunits=3^3, randomize=~block/UNITS)
print(mydesign)
# DUMMY ANALYSIS
# Here we omit two-factor interactions from the model, so they are
# confounded with the residuals (but not with ABCD main effects)
set.seed(123)
mydesigndata <- mydesign@design
mydesigndata$Y <- runif(27)
mydesign.aov <- aov(Y ~ block + A + B + C + D, data=mydesigndata)
summary(mydesign.aov)
# 2. STEP-BY-STEP GENERATION, USING planor.designkey
F0 <- planor.factors(factors=c( "block", LETTERS[1:4]), nlevels=rep(3,5),
block=~block)
M0 <- planor.model(model=~block+(A+B+C+D)^2, estimate=~A+B+C+D)
K0 <- planor.designkey(factors=F0, model=M0, nunits=3^3, max.sol=2)
summary(K0)
mydesign.S4 <- planor.design(key=K0, select=2)
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.