Prev: How to change color and Line property in Pwelch
Next: Out of Windows Resources: Allocation of bitmap failed. Disabling
From: carlas on 19 Nov 2009 06:13
If what you say is true “that this result is off because of general statistics” then limit of negE when the number of samples go to infinity should go to zero. This is not the case, since it explodes, so there should be something wrong with the normalization.


I = randn(10000000,1);
p = hist(I,1000000);
% remove zero entries in p
p(p==0) = [];
% normalize p so that sum(p) is one.
p = p ./ numel(I);
Eg = -sum(p.*log(p));

I = randn(10000000,1);
p = hist(I,1000000);
% remove zero entries in p
p(p==0) = [];
% normalize p so that sum(p) is one.
p = p ./ numel(I);
Et = -sum(p.*log(p));


negE=Et-Eg
 | 
Pages: 1
Prev: How to change color and Line property in Pwelch
Next: Out of Windows Resources: Allocation of bitmap failed. Disabling