And it’s of interest – does evolutionary nature of the search algorithm has any amenity over simple random search in terms of fitness function evaluations count (=computation time) or quality of solution?
Good day! thanks for interesting program. Can anyone please explain me why, having entropy(rand(10000,1))=0 (it's ok),
it gives entropy(sin(rand(10000,1)))=0.9978, BUT entropy(cos(rand(10000,1)))=0
I believe entropy(sin(rand(10000,1))) should also equal 0... or not? with cos(rand) or sin(rand) signals we definitely have less information than with pure rand signal?
For those who can't run the demo file because of 'undefined function' error. You need to:
1) Run this command: list = dir('*.cpp');
to get the list of files.
2) For all the files in list, change log(2) to log(2.0). You can get all the files' names by:
2) Run 'makeosmex.m'
Now this should works like a dream.
I got following error, on running demo_mi.m
Undefined function 'estpab' for input arguments of type 'double'.
Error in mutualinfo (line 21)
[p12, p1, p2] = estpab(vec1,vec2);
Error in demo_mi (line 25)
What's wrong ?
I also get the problem "Undefined function 'estpa' for input arguments of type 'double'". I find the estpa is cpp file which is a C++ file. I use Win8 62 bit application and Matlab 2011b. What's wrong with it?
when the range of the features vectors is large, let's say 1-10000 estpab() calculates probibilty density with the same length (10000 in this case). Now if both feature vectors have a large range, lets say the other one is 1-10000 too, then the joint pdf computed by estpab is a 10000 x 10000 matrix and the computation if MI takes very long time. Is there a way to modify the algorithm in order to avoid this problem?