Hi!
I didn't see the toy data, my bad.
I guess David's solution fixed the problem.
Ivan
Le 8/26/2010 16:10, Bryan Hanson a écrit :
> Hi Ivan, there is toy data given in the original post.
>
> The object passed to aov is a matrix, it is called scores, and it is passed
> via the formula that is created in the sense that the formula specifies
> objects in the calling environment. At least that is what is supposed to
> happen.
>
> I've not been able to find an example of manova using a data frame - aov,
> yes. I think that is because the response vector in manova must be a
> matrix, and you can't put a matrix in a data frame as a separate object.
> You could put the columns of the matrix into the data frame, but then it is
> not seen as a single entity, it's seen as separate columns, and specifying
> this in the formula looks impossible as I understand it.
>
> But, obviously there is something I don't understand! Thanks, Bryan
>
>
> On 8/26/10 9:50 AM, "Ivan Calandra"<ivan.calan...@uni-hamburg.de> wrote:
>
>> Hi,
>>
>> First, some toy data would have helped a lot, and that can explain why
>> you got so few answers...
>>
>> Second, I have maybe some thoughts about it, not sure if this can help.
>>
>> The error tells you that the data passed to aov is not a matrix. In your
>> function, you don't specify explicitly the data object to be used, that
>> might be a problem. And that object has to be a matrix, it might not be.
>> Maybe something like:
>>
>> form<- as.formula(paste("pcs", "~", paste(fac, collapse = "*")),
>> ...
>> if (length(pcs) == 1) out<- aov(formula = form, data=score.matrix)
>>
>> That would work if I uderstood what you're doing (hum...) and if both "pcs"
>> and "fac" are parts of "score.matrix". But I've always done aov from a
>> dataframe, which makes sense if you have both numerical variables and
>> factors.
>> How can you do it from a matrix? Maybe my understanding is limited.
>>
>> As I said, just some thoughts...
>> HTH,
>> Ivan
>>
>>
>> Le 8/26/2010 14:47, Bryan Hanson a écrit :
>>> Hello Again Gurus and Lurkers:
>>>
>>> I¹m trying to build a very user-friendly function which does aov without
>>> having the user type in a formula (which would be tedious in this case).
>>> The idea is to take the response from a PCA score matrix, and the factors
>>> from a list. A simple example is the function given below, along with test
>>> data and a sample call to the function.
>>>
>>> I'm certainly having trouble understanding the proper ways to work with
>>> formulas and related items, but I think what I do in the function should
>>> work (it's built on ideas dug out of the archives). However, when the data
>>> is passed to aov (directly or via manova), something in the bowels of aov
>>> complains with the following error:
>>>
>>> Error in model.frame.default(formula = form, drop.unused.levels = TRUE) :
>>> object is not a matrix
>>>
>>> To me, the formula looks legitimate, and the variables in the formula are
>>> all in the environment (I think: The way I am doing this is basically that
>>> described in ?summary.manova where only a formula is passed, no data
>>> argument). Based upon reading the archives, the problem might arise with
>>> one of the deparse statements in aov, but I can't resolve it. It might also
>>> be one of scoping/environment, but again, this is only an idea.
>>>
>>> TIA for any assistance. Bryan
>>> *************
>>> Bryan Hanson
>>> Professor of Chemistry& Biochemistry
>>> DePauw University, Greencastle IN USA
>>>
>>>
>>> hypTestScores<-
>>> function(mylist, score.matrix, pcs = 1:3, fac = NULL, ...) {
>>>
>>> scores<- score.matrix[,pcs]
>>> # str(scores) # looks correct to me
>>> form<- as.formula(paste("scores", "~", paste(fac, collapse = "*")),
>>> env = parent.frame())
>>> # str(form) # looks correct to me
>>> attach(mylist)
>>> if (length(pcs)> 1) out<- manova(formula = form, ...)
>>> if (length(pcs) == 1) out<- aov(formula = form, ...)
>>> print(summary(out))
>>> detach(mylist)
>>> invisible(out)
>>> }
>>>
>>> # test data
>>> td1<- matrix(rnorm(50), ncol = 5) # like PCA scores
>>> td2<- list(
>>> f1 = as.factor(sample(c("A", "B"), 10, replace = TRUE)),
>>> f2 = as.factor(sample(c("C", "D"), 10, replace = TRUE)))
>>>
>>> # test call
>>> hypTestScores(mylist = td2, score.matrix = td1,
>>> fac = c("f1", "f2"))
>>> detach("mylist") # needed if there is an error
>>>
>>>> sessionInfo()
>>> R version 2.11.0 (2010-04-22)
>>> x86_64-apple-darwin9.8.0
>>>
>>> locale:
>>> [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
>>>
>>> attached base packages:
>>> [1] datasets tools grid graphics grDevices utils stats
>>> [8] methods base
>>>
>>> other attached packages:
>>> [1] faraway_1.0.4 GGally_0.2 xtable_1.5-6
>>> [4] mvbutils_2.5.1 ggplot2_0.8.8 proto_0.3-8
>>> [7] reshape_0.8.3 ChemoSpec_1.45 R.utils_1.4.0
>>> [10] R.oo_1.7.2 R.methodsS3_1.2.0 rgl_0.91
>>> [13] lattice_0.18-5 mvoutlier_1.4 plyr_1.0.3
>>> [16] RColorBrewer_1.0-2 chemometrics_0.8 som_0.3-5
>>> [19] robustbase_0.5-0-1 rpart_3.1-46 pls_2.1-0
>>> [22] pcaPP_1.8-1 mvtnorm_0.9-9 nnet_7.3-1
>>> [25] mclust_3.4.4 MASS_7.3-5 lars_0.9-7
>>> [28] e1071_1.5-23 class_7.3-2
>>>
>>> ______________________________________________
>>> 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.
>>>
>
>
--
Ivan CALANDRA
PhD Student
University of Hamburg
Biozentrum Grindel und Zoologisches Museum
Abt. Säugetiere
Martin-Luther-King-Platz 3
D-20146 Hamburg, GERMANY
+49(0)40 42838 6231
ivan.calan...@uni-hamburg.de
**********
http://www.for771.uni-bonn.de
http://webapp5.rrz.uni-hamburg.de/mammals/eng/mitarbeiter.php
[[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.
