using histogram in a loop with hold on and off

I am tyring to understand the use of the histogram in the following way.

I have a matrix of size (M x N). Each column represents a curve.

I calculate a mean for all the curves using mean_matrix = mean(matrix,2)

Plotting mean_matrix shows that the mean is as expected.

Next I subtract each curve from the mean value giving me a difference matrix Diff(M x N)

I then use histogram to plot Diff(M x N) so it is > hist(Diff, # of bins)

This give me a histogram with different colours, I think the different colours

are for each column of the Diff matrix ? I then took a look at using

 a loop so I did >for i=1,index; hist(Diff(:,i),# of bins);end; and this

gave me a different looking graph; next I included hold on and again

I got a completely looking curve. Can someone help me understand what I did?

I would like to do is to look at the histogram curves in a 3D plot

and see the variation better.

Answer 1

[Michael Haderlein]

I try to understand and to explain. Let's take a matrix, maybe the peaks-matrix (49x49 by default) and reshape it:

matrix=peaks; %49x49
matrix2=reshape(matrix,numel(matrix)/7,7); %343x7
hist(matrix2) %histogram with 7 bars each bin, one for each column of the matrix
figure, axes, hold all
for cnt=1:7
hist(matrix2(:,cnt))
end %7 histograms stacked upon each other -> difficult to see anything
h3=histc(matrix2,linspace(minv,maxv,21),1); %calculate a histogram for each column
bar3(h3) %plot the histogram for each column

Hope that helped,

best regards,

Michael

[Philip]

Michael, thanks for the feedback. Can you help me understand the use of the

linspace. I extracted out the bin counts and the bin values using the following

 for n=1:index;

    [bincounts,bins] = hist(Diff(:,index),No_bins);

    bin_count = transpose(bincounts);
  bin_count_t(:,n) = bin_count;
    bin_value = transpose(bins);
end

The matrix bin_count_t (M x N) now contains the number of occurance of the

difference between the mean and each curve. bin_value is (M x 1) and

contains the values of each bin.

I then do a bar (bin_value, bin_count_t) again the stacking occurs as you

noted. Now since I already have the bin count and bin values extracted

how do I use linspace to separate out the curves ?, I experimented with

bar3 but not making progress, thanks inadvance for you feedback if you have time.

Regards

Philip

Answer 2

[Michael Haderlein]

I see I forgot to post also the line where minv,maxv is created:

minv=min(matrix2(:));maxv=max(matrix2(:));

What happens, if you execute the above written just using your Diff matrix instead of my matrix2?