| array indexing question [message #31442] |
Thu, 11 July 2002 15:57  |
mmiller3
Messages: 81
Registered: January 2002
|
Member |
|
|
I'm doing some calculations to calculate mutual information and I
have a question about the most efficient/slickest way to
calculate the sum. I wonder if there is a loop-free way to do
this.
I have three arrays, pa, pb and pab. pa and pb are both 1D
arrays of length N and pab is a 2D NxN array. I want to
calculate the sum of pab[i,j]*alog(pab[i,j])/pa[i]/pb[j]. I know
that I can do things like total( pa * alog(pa) ) when I'm dealing
with a single array. Any suggestions for how to do my first sum
most efficiently?
Mike
--
Michael A. Miller mmiller3@iupui.edu
Imaging Sciences, Department of Radiology, IU School of Medicine
|
|
|
|
| Re: Array indexing [message #31548 is a reply to message #31442] |
Tue, 16 July 2002 01:50  |
marc schellens[1]
Messages: 183
Registered: January 2000
|
Senior Member |
|
|
Andre Kyme wrote:
> Hi everyone,
>
> I have a 2D image and want to index a bunch of (x,y) pairs.
> X is the set of x coords and Y the set of y coords. Say I want
> to set all (x,y) pairs to 200:
>
> image[X,Y]=200
>
> This is fine. But if image has multiple slices and I do:
>
> image[X,Y,0]=200
>
> all possible combinations of points in X and Y get set to 200,
> whereas I only want corresponding values in X and Y set.
> Is there some array notation I'm missing that enables this?
All possible combinations you get when the indexing arrays are of
different size.
So here you want to write:
Z=lonarr(n_elements(X))
image[X,Y,Z]=200
cheers,
:-) marc
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
