Hello.
I would like to estimate the correlation coefficient
from two samples with Bootstrapping using the
R-function sample().
The problem is, that I have to sample pairwise. For
example if I have got two time series and I draw from
the first series the value from 1912 I need the value
from 1912 from the second sample, too.
Example:
Imagine that a and b are two time series with returns
for example:
a <- c(1,2,3,4,5,6,7,8,9,10)
b <- c(1,1,56,3,6,6,6,7,2,10)
a.sample <- numeric(10)
b.sample <- numeric(10)
boot.cor.a.b <- numeric(1000)
for (i in 1:1000)
{
for (j in 1:10)
{
a.sample[j] <- sample(a,1,replace=TRUE)
b.sample[j] <- sample(b,1,replace=TRUE)
}
boot.cor.a.b[i] <- cor(a,b)
}
The problem here is, that the sampling is independent
from each other.
So how do I have to change the R-code to get the
pairwise sampling mentioned above?
I hope you can help me.
Sincerely
Klein.
________
Try this:
a <- c(1,2,3,4,5,6,7,8,9,10)
b <- c(1,1,56,3,6,6,6,7,2,10)
n <- length(a)
boot.cor.a.b <- replicate( 1000, {tmp <- sample(n, replace=TRUE);
cor(a[tmp],b[tmp]) } )
hist(boot.cor.a.b)
abline( v=c( mean(boot.cor.a.b), median(boot.cor.a.b) ),
col=c('blue','green'))
Hope this helps,
--
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at intermountainmail.org
(801) 408-8111
> -----Original Message-----
> From: r-help-bounces at r-project.org
> [mailto:r-help-bounces at r-project.org] On Behalf Of Andreas Klein
> Sent: Thursday, September 20, 2007 11:59 PM
> To: r-help at r-project.org
> Subject: [R] Estimate correlation with bootstrap
>
> Hello.
>
> I would like to estimate the correlation coefficient from two
> samples with Bootstrapping using the R-function sample().
>
> The problem is, that I have to sample pairwise. For example
> if I have got two time series and I draw from the first
> series the value from 1912 I need the value from 1912 from
> the second sample, too.
>
> Example:
>
> Imagine that a and b are two time series with returns for example:
>
> a <- c(1,2,3,4,5,6,7,8,9,10)
> b <- c(1,1,56,3,6,6,6,7,2,10)
>
> a.sample <- numeric(10)
> b.sample <- numeric(10)
> boot.cor.a.b <- numeric(1000)
>
> for (i in 1:1000)
>
> {
>
> for (j in 1:10)
>
> {
>
> a.sample[j] <- sample(a,1,replace=TRUE)
> b.sample[j] <- sample(b,1,replace=TRUE)
>
> }
>
> boot.cor.a.b[i] <- cor(a,b)
>
> }
>
> The problem here is, that the sampling is independent from each other.
>
> So how do I have to change the R-code to get the pairwise
> sampling mentioned above?
>
> I hope you can help me.
>
> Sincerely
> Klein.
>
>
> ________
>
> ______________________________________________
> R-help at 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.
>
Hi Andreas,
Simply use one call to the sample function.
Try this:
a <- c(1,2,3,4,5,6,7,8,9,10)
b <- c(1,1,56,3,6,6,6,7,2,10)
boot.cor.a.b <- numeric(1000)
for (i in 1:1000){
x=sample(10,replace=T)
boot.cor.a.b[i]=cor(a[x],b[x])
}
Note that in R you don't need to initialize the variables like in other
languages.
Another option is to use the boot() function directly. See ?boot.
Julian
Andreas Klein wrote:> Hello.
>
> I would like to estimate the correlation coefficient
> from two samples with Bootstrapping using the
> R-function sample().
>
> The problem is, that I have to sample pairwise. For
> example if I have got two time series and I draw from
> the first series the value from 1912 I need the value
> from 1912 from the second sample, too.
>
> Example:
>
> Imagine that a and b are two time series with returns
> for example:
>
> a <- c(1,2,3,4,5,6,7,8,9,10)
> b <- c(1,1,56,3,6,6,6,7,2,10)
>
> a.sample <- numeric(10)
> b.sample <- numeric(10)
> boot.cor.a.b <- numeric(1000)
>
> for (i in 1:1000)
>
> {
>
> for (j in 1:10)
>
> {
>
> a.sample[j] <- sample(a,1,replace=TRUE)
> b.sample[j] <- sample(b,1,replace=TRUE)
>
> }
>
> boot.cor.a.b[i] <- cor(a,b)
>
> }
>
> The problem here is, that the sampling is independent
> from each other.
>
> So how do I have to change the R-code to get the
> pairwise sampling mentioned above?
>
> I hope you can help me.
>
> Sincerely
> Klein.
>
>
> ________
>
> ______________________________________________
> R-help at 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.
>
>
--
Julian M. Burgos
Fisheries Acoustics Research Lab
School of Aquatic and Fishery Science
University of Washington
1122 NE Boat Street
Seattle, WA 98105
Phone: 206-221-6864
Andreas,
To do pairwise bootstrap sampling, try
for (i in 1:1000) {
ind<-sample(1:10,10,replace=TRUE)
boot.com.ab[i]<-cor(a[ind],b[ind])
}
Regards,
-Cody
Cody Hamilton
Edwards Lifesciences
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On Behalf Of Andreas Klein
Sent: Thursday, September 20, 2007 10:59 PM
To: r-help at r-project.org
Subject: [R] Estimate correlation with bootstrap
Hello.
I would like to estimate the correlation coefficient
from two samples with Bootstrapping using the
R-function sample().
The problem is, that I have to sample pairwise. For
example if I have got two time series and I draw from
the first series the value from 1912 I need the value
from 1912 from the second sample, too.
Example:
Imagine that a and b are two time series with returns
for example:
a <- c(1,2,3,4,5,6,7,8,9,10)
b <- c(1,1,56,3,6,6,6,7,2,10)
a.sample <- numeric(10)
b.sample <- numeric(10)
boot.cor.a.b <- numeric(1000)
for (i in 1:1000)
{
for (j in 1:10)
{
a.sample[j] <- sample(a,1,replace=TRUE)
b.sample[j] <- sample(b,1,replace=TRUE)
}
boot.cor.a.b[i] <- cor(a,b)
}
The problem here is, that the sampling is independent
from each other.
So how do I have to change the R-code to get the
pairwise sampling mentioned above?
I hope you can help me.
Sincerely
Klein.
________
______________________________________________
R-help at 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.