bootci usage for a newbie to bootstrap methods
in [Matlab]
|
Prev: Exit Debug Mode clears breakpoints (sometimes)
Next: again: problem with using function to flip any matrix and give it a new name
From: Ben Trettel on 28 Feb 2010 21:20 I need to make a confidence interval using bootstrap methods for some non-parametric data. The remainder of the statistical analysis I have done in MATLAB, so I want to do this part in MATLAB too to save some time. I know bootci is what I want to use in MATLAB. Unfortunately, my knowledge of bootstrap methods ends there. After a few hours of searching I am yet to find an adequate explanation of the command. I'm knowledgeable about relatively basic statistics, but this is completely foreign. Could someone explain roughly what I have to do when I have a certain set of data? I assume this is a simple task when one knows what each parameter means. If not, let me know. Ben Trettel
From: Peter Perkins on 1 Mar 2010 10:51
On 2/28/2010 9:20 PM, Ben Trettel wrote: > After a few hours of searching I am yet to find an adequate > explanation of the command. I'm knowledgeable about relatively basic > statistics, but this is completely foreign. > > Could someone explain roughly what I have to do when I have a certain > set of data? I assume this is a simple task when one knows what each > parameter means. If not, let me know. Ben, other than pointing you to the example in the help, it's hard to give specific help without knowing what it is you're doing and what you're having trouble with. In general, you'd call BOOTCI by specifying 1) the number of bootstrap iterations you'd like to use (1000, say) 2) a function that computes whatever statistic you're bootstrapping (MEAN, say), and 3) your data, in the form of a column vector or vectors There are a lot of other optional arguments, but those are the basic ones. The example in the help might be a starting point. Hope this helps. |