|
»
»
« Return to Thread: applying cor.test to a (m, n) matrix
Re: applying cor.test to a (m, n) matrix
by Kenn Konstabel
::
Rate this Message:
I've used the following function (I wrote it some time ago so I don't
remember any more why I needed it, but I checked and it still works). I
don't think you can get rid of for loops altogether: if you look at the code
of apply, you'll see some there too.
The argument STATS specifies those components of a function's output you
want to extract (if the component you're asking for has length >1 then
you'll get the first item). To extract ci's you will need an intermediate
function:
cor.test2 <- function(...) {
foo<-cor.test(...)
ci<-foo$conf.int
foo$ci1<-ci[1]
foo$ci2<-ci[2]
foo
}
Then rrapply(my.data, FUN=cor.test2, STATS=c("estimate", "p.value", "ci1",
"ci2")) should do what you want.
rrapply <-
function(x, FUN=t.test, STATS=c("statistic", "p.value"), ...){
# use FUN=cor.test, STATS=c("estimate", "p.value") for matrix of
correlations and p-values
K <- ncol(x)
RESU <- list()
my.fun <- FUN
fuf<-my.fun(x[,2],x[,1], ...)
if(!is.list(fuf)) {
STATS <- "A"
my.fun <- function(x,y,...) list(A=FUN(x,y,...))
}
neimz <- colnames(x)
for(i in 1:length(STATS)) {
fof <- STATS[i]
RESU[[fof]] <- matrix(NA, ncol=K, nrow=K)
colnames(RESU[[fof]]) <- neimz
rownames(RESU[[fof]]) <- neimz
names(RESU)[i] <- fof
}
for(i in 2:K) for(j in 1:(i-1)) {
foo<-my.fun(x[,i],x[,j], ...)
for(h in 1:length(STATS)) {
fof <- STATS[h]
RESU[[fof]][i,j]<- foo[[fof]]
}
}
if(is.list(fuf)) RESU else RESU$A
}
On Fri, May 9, 2008 at 10:12 AM, Dimitris Rizopoulos <
dimitris.rizopoulos@...> wrote:
> have a look also at function rcor.test() from package ltm.
>
> Best,
> Dimitris
>
> ----
> Dimitris Rizopoulos
> Biostatistical Centre
> School of Public Health
> Catholic University of Leuven
>
> Address: Kapucijnenvoer 35, Leuven, Belgium
> Tel: +32/(0)16/336899
> Fax: +32/(0)16/337015
> Web: http://med.kuleuven.be/biostat/
> http://www.student.kuleuven.be/~m0390867/dimitris.htm<http://www.student.kuleuven.be/%7Em0390867/dimitris.htm>
>
>
> ----- Original Message ----- From: "Monica Pisica" <pisicandru@...
> >
> To: <r-help@...>
> Sent: Thursday, May 08, 2008 9:05 PM
> Subject: [R] applying cor.test to a (m, n) matrix
>
>
>
>
>> Hi everybody,
>>
>> I would like to apply cor.test to a matrix with m rows and n columns and
>> get the results in a list of matrices , one matrix for p.val, one for the
>> statistic, one for the correlation and 2 for upper and lower confidence
>> intervals, something analog with cor() applied to a matrix.
>>
>> I have done my own function to get a matrix of p.values and i suppose i
>> can build similar functions for all the others. But i have used for loops
>> and i wonder if there is any way to actually use one of the functions from
>> the "apply" family to do this in a quicker way.
>>
>> Here is my little function:
>>
>> cor.pval <- function(x, method = c("pearson", "kendal", "spearman"),
>> digit=8) {
>> n <- dim(x)[2]
>> pval <- matrix(paste(rep("c", n*n), seq(1,n*n), sep = ""), n, n, byrow =
>> T)
>> for (i in 1:n) {
>> for (j in 1:n){
>> pval[i, j] <- cor.test(x[,i], x[,j], method = method)$p.value
>> }
>> }
>> pval <- matrix(round(as.numeric(pval),digit), n, n, byrow = T)
>> rownames(pval) <- colnames(x)
>> colnames(pval) <- colnames(x)
>> return(pval)
>> }
>>
>> Thanks for any input,
>>
>> Monica
>>
>>
>>
>>
>>
>>
>>
>> _________________________________________________________________
>>
>>
>> esh_messenger_052008
>> ______________________________________________
>> R-help@... mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>>
>
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
>
>
> ______________________________________________
> R-help@... mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
[[alternative HTML version deleted]]
______________________________________________
R-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
remember any more why I needed it, but I checked and it still works). I
don't think you can get rid of for loops altogether: if you look at the code
of apply, you'll see some there too.
The argument STATS specifies those components of a function's output you
want to extract (if the component you're asking for has length >1 then
you'll get the first item). To extract ci's you will need an intermediate
function:
cor.test2 <- function(...) {
foo<-cor.test(...)
ci<-foo$conf.int
foo$ci1<-ci[1]
foo$ci2<-ci[2]
foo
}
Then rrapply(my.data, FUN=cor.test2, STATS=c("estimate", "p.value", "ci1",
"ci2")) should do what you want.
rrapply <-
function(x, FUN=t.test, STATS=c("statistic", "p.value"), ...){
# use FUN=cor.test, STATS=c("estimate", "p.value") for matrix of
correlations and p-values
K <- ncol(x)
RESU <- list()
my.fun <- FUN
fuf<-my.fun(x[,2],x[,1], ...)
if(!is.list(fuf)) {
STATS <- "A"
my.fun <- function(x,y,...) list(A=FUN(x,y,...))
}
neimz <- colnames(x)
for(i in 1:length(STATS)) {
fof <- STATS[i]
RESU[[fof]] <- matrix(NA, ncol=K, nrow=K)
colnames(RESU[[fof]]) <- neimz
rownames(RESU[[fof]]) <- neimz
names(RESU)[i] <- fof
}
for(i in 2:K) for(j in 1:(i-1)) {
foo<-my.fun(x[,i],x[,j], ...)
for(h in 1:length(STATS)) {
fof <- STATS[h]
RESU[[fof]][i,j]<- foo[[fof]]
}
}
if(is.list(fuf)) RESU else RESU$A
}
On Fri, May 9, 2008 at 10:12 AM, Dimitris Rizopoulos <
dimitris.rizopoulos@...> wrote:
> have a look also at function rcor.test() from package ltm.
>
> Best,
> Dimitris
>
> ----
> Dimitris Rizopoulos
> Biostatistical Centre
> School of Public Health
> Catholic University of Leuven
>
> Address: Kapucijnenvoer 35, Leuven, Belgium
> Tel: +32/(0)16/336899
> Fax: +32/(0)16/337015
> Web: http://med.kuleuven.be/biostat/
> http://www.student.kuleuven.be/~m0390867/dimitris.htm<http://www.student.kuleuven.be/%7Em0390867/dimitris.htm>
>
>
> ----- Original Message ----- From: "Monica Pisica" <pisicandru@...
> >
> To: <r-help@...>
> Sent: Thursday, May 08, 2008 9:05 PM
> Subject: [R] applying cor.test to a (m, n) matrix
>
>
>
>
>> Hi everybody,
>>
>> I would like to apply cor.test to a matrix with m rows and n columns and
>> get the results in a list of matrices , one matrix for p.val, one for the
>> statistic, one for the correlation and 2 for upper and lower confidence
>> intervals, something analog with cor() applied to a matrix.
>>
>> I have done my own function to get a matrix of p.values and i suppose i
>> can build similar functions for all the others. But i have used for loops
>> and i wonder if there is any way to actually use one of the functions from
>> the "apply" family to do this in a quicker way.
>>
>> Here is my little function:
>>
>> cor.pval <- function(x, method = c("pearson", "kendal", "spearman"),
>> digit=8) {
>> n <- dim(x)[2]
>> pval <- matrix(paste(rep("c", n*n), seq(1,n*n), sep = ""), n, n, byrow =
>> T)
>> for (i in 1:n) {
>> for (j in 1:n){
>> pval[i, j] <- cor.test(x[,i], x[,j], method = method)$p.value
>> }
>> }
>> pval <- matrix(round(as.numeric(pval),digit), n, n, byrow = T)
>> rownames(pval) <- colnames(x)
>> colnames(pval) <- colnames(x)
>> return(pval)
>> }
>>
>> Thanks for any input,
>>
>> Monica
>>
>>
>>
>>
>>
>>
>>
>> _________________________________________________________________
>>
>>
>> esh_messenger_052008
>> ______________________________________________
>> R-help@... 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.
>>
>>
>
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
>
>
> ______________________________________________
> R-help@... 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.
>
[[alternative HTML version deleted]]
______________________________________________
R-help@... 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.
« Return to Thread: applying cor.test to a (m, n) matrix
| Free Forum Powered by Nabble | Forum Help |