MATLAB Answers

0

Counting number of 1's in different bins

Asked by MiauMiau on 8 May 2015
Latest activity Commented on by MiauMiau on 8 May 2015
Hi
I have again a question regarding bins:
Again an array A:
A = [100 150 190 200 250 300 350 370 390 400 400]
and a second array B which saves if a stimuli with the intensity as defined in A was recognized by a subject or not:
B = [ 0 1 0 0 1 1 1 0 1 0 1]
edges = linspace(min(A),max(A),4);
First I bin A in the bins [100,200), [200,300) and [300,400] as follows:
temp = histc(A, edges);
counts = temp(1:end-1);
counts(end) = counts(end)+ temp(end);
Now I want to know for each bin, how many 1's it contains (as stored in B). How do I do that? I started to code with a for loop but the code became quite complicated, so maybe my approach was wrong to start with..

  2 Comments

Adam
on 8 May 2015
If possible can you post the exact definition of stim_durations and edges for that code?
It is a lot easier to try to help if I (or anyone) can just quickly put your code on our own command line or script without having to interpret it and work out what is supposed to be in a variable.
histc does come with a second output argument:
[bincounts,ind]= histc(___)
ind, being the index of the bin into which each element was placed. Using that to apply to your 'B' array above should quickly yield the answer you need.
As an aside though the Matlab documentation recommends to use histcounts instesad of histc though I don't know which matlab version you are running or which version histcounts was introduced as the recommended alternative.
I have unfortunately an older Matlab version - and since I have to change the output of histc (see last edge) it does not exactly output the indices I need (thought is still useful I think). I corrected my question now. Thanks though!

Sign in to comment.

3 Answers

Answer by Michael Haderlein on 8 May 2015
Edited by Michael Haderlein on 8 May 2015

histc can have two output parameters:
[temp,ind] = histc(A,[100,200,300,400]);
Use the second one as input parameter for accumarray:
>> accumarray(ind',B',[],@sum)
ans =
1
1
3
1

  0 Comments

Sign in to comment.


Answer by Purushottama Rao on 8 May 2015

For a binary vector B
K= sum(B==1)
will return no of 1s

  1 Comment

that is not the question. the question is number of 1s for each bin

Sign in to comment.


Answer by Andrei Bobrov
on 8 May 2015

A = [100 150 190 200 250 300 350 370 390 400 400];
B = [ 0 1 0 0 1 1 1 0 1 0 1];
[~,ii] = histc(A,[100 200 300 inf]);
out = accumarray(ii(:),B(:));

  0 Comments

Sign in to comment.