bootstrapping

Hello,



I was able to bootstrap 1000 times, 20 times (20 columns of 1000  
values).  Is there a way to get mean for each column?

Thanks!



Aaron

Date: Sat, 16 Jan 2010 00:04:59 +0100
From: Stephan.Kolassa at gmx.de
To: aaron.foley at students.tamuk.edu
CC: r-help at r-project.org
Subject: Re: [R] bootstrapping

Hi Aaron,

try the argument "statistic=mean". Then boot() will give you the mean
turn angle in your actual data (which appears to be 6 degrees,  
judging
from what you write), as well as the means of the bootstrapped data.
Then you can get (nonparametric) bootstrap CIs by
quantile(boot$t,probs=c(.025,.975)). As far as I can see, there is
really no need to look at sd().

A more interesting question would be how to deal with the fact that
-180=+180, there may be something to think about here...

HTH,
Stephan


aaron.foley at students.tamuk.edu schrieb:

Hi All,



I'm new to R so please bear with me. I have a dataset with 337  
turn angles ranging from -180 to 180 degrees. I need to bootstrap  
(sample with replacement) 1,000 times to create expected average  
turn angle with 95% CIs. The code is pretty straightforward (<- 
boot(data =, statistic = ,R =)) but I am unsure how to input my  
observed mean (6 degrees) and standard deviation (66 degrees) into  
the statistic component. I realize there is a 'function' code but  
I can't seem to carry the results over to the 'boot' code.



Thanks,

Aaron M. Foley
PhD Candidate
Caesar Kleberg Wildlife Research Institute
Texas A&M University - Kingsville
Cousins Hall, Room 201
Kingsville, TX 78363




[[alternative HTML version deleted]]

		 	   		
	[[alternative HTML version deleted]]