Dear R users,
I have used the AlgDesign package to construct a D-optimal exp. design from
a file containing a set of allowed runs. The code for the optFederov call is
below:
#Using optFederov ------------------------------
library(AlgDesign)
options(contrasts=c("contr.sum","contr.poly"))
file<- read.table("C:/Upstream/Designs/Vodacom Tanzania phase1
grammar/JMP_allowed.txt",sep="\t",header=TRUE,strip.white=TRUE)
formula<-as.formula("~structure+intro + intro/intro1 + concept +
concept/concept1 + prize")
design<-optFederov(formula, data=file, nTrials=48, evaluateI=T,
criterion="D", maxIteration=10000, nRepeats=300,
approximate=F,args=T)
#------------------------------------------------------------------------------------------------------------------------------------------------------------------------
I also use the Rcmdr interface at time for ease, when using the Rcmdr
interface is used the same data file, formula and desired number of runs. I
also set the seed to that of the optFederov call above. I was expected the
same design to be generated. However, when I examine the balance of the
generated design, Dopt.design and optFederov yield different designs. Given
that Dopt.design is a DoE wrapper for optFerderov I am confused as to why
this is happening.
#Using R-commander
Dataset.Dopt <- Dopt.design(48, data= file, formula,
nRepeat= 300 ,randomize= TRUE, seed=403)
Apologies if this has a simple answer, I have consulted the R-help archives,
but have not found an answer as yet. If the forum could offer reasons as to
why this is happening I would be most grateful.
Regards
M. Griffiths
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
