Using hist3 and sorting

7 views (last 30 days)
Michael
Michael on 20 Jan 2016
Edited: Stephen23 on 22 Jan 2016
I'm simulating event positions and times, and histogramming then in the XY plane. I want to find for each event, the time difference to the next event in the same bin. I feel liek this should be a ~4 liner or so...
posXY = rand(nEv,2)
times = rand(nEv,1)
posBin = [2 2]; % put the events on a 2x2 mesh in X-Y plane
[hitMatrix, ~] = hist3(posXY, posBin);
then I want to find (for each event, in a given bin--I made four bins!) the time the next event occurs in that position bin. So i thought about sorting the events by their times, and using the indices to pick out the correct events that also occur in the same bin... making a logical and looping over events.. but I can't quite nail the whole sequence down. Do you have any tips?
I appreciate your input, Michael B.

Accepted Answer

Stephen23
Stephen23 on 20 Jan 2016
Edited: Stephen23 on 20 Jan 2016
The trick is to get the indices from the histogram, but unfortunately hist3 does not return the indices of the input values. Introduced in 2015b is the function histcounts2, which does return these indices. However I do not have 2015b, so I found histcn on FEX, which seems to do what you require.
I tried it like this, is three lines okay?:
nEv = 10;
posXY = rand(nEv,2);
times = rand(nEv,1);
%
[idx,~,~,loc] = histcn(posXY,2,2);
A = accumarray(loc,times,[],@(n){n});
B = cellfun(@(v)diff(sort(v)),A,'UniformOutput',false)
Where B is a cell array where the cells correspond to the bins of the histogram and contain the differences in the sorted times values corresponding to the X and Y values that were sorted into the corresponding bin. The variable A contains the time values themselves, so you can check this by hand if you want.
  4 Comments
Michael
Michael on 21 Jan 2016
Thanks for all your help! I'm still parsing the original solution but your input is very much appreciated (I never looked at length help--thinking I had mastered its use... jeez I feel like a newb now) Again, your time and help are much appreciated!

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!