I am now writing an R code to do the systematic permutation test and the
random permutation test.
And I encountered a couple of problems which I cannot figure them
out. At first, whenever I type "permtest" in my R2.6.1, it always
shows:"there is no such a function", I tried again and again and
eventually
I found I have to use "library (BHH2)" before I use
"permtest", though it
make me very glad, the question is It just works for two samples situation,
but does not work for 3 samples at all.
Well, What I want to do is to use both the systematic permutation and the
random permutation to test H0:µA= µB =µC
Where, say A<-(1,2), B(3, 4, 5), C<-(6, 7, 8, 9), What I want to find is
just the F-value for one-way ANOVA.
This is my R code and result when I am just use A and B> library(BHH2)
>
> A <- c(1,2)
> B <- c(3,4,5)
>
> permtest(A,B)
N t.obs t-Dist:P(>t) PermDist:P(>t) F.obs
10.0000000 -3.0000000 0.9711656 0.9000000 0.5000000
F-Dist:P(>F) PermDist:P(>F)
0.5527864 0.5000000>
And you see it gives me a lot of information, and I just know N=10 is just
the number of possible permutations. I do not know whether it should be
systematic permutation test or the random permutation test and I do not know
which should be the F for one-way ANOVA.
So I want to ask you Dr.Ross Ihaka could you give me some hints or example
on how to do the systematic permutation test and the random permutation test
when we have three samples and provide the F for one-way ANOVA.
I know this email may will distract you, but I really need your help.
Thanks you very much
Tony
Feb 26, 2008
[[alternative HTML version deleted]]
