Pairwaise Mutual Information Calculation

3 次查看（过去 30 天）

Marimuthu Ananthavelu 2018-12-23

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/437022-pairwaise-mutual-information-calculation

编辑： Marimuthu Ananthavelu 2018-12-24

data.csv

I am trying to estimate the Mutual Information for pairwise combinations for the whole matrix (which is an EEG data). I use the functions defined here link. And I use following code to estimate this pairwise.

Data=csvread('data.csv'); % loads the EEG data.
% MI for pair of recordings in EEG
n=size(Data,1);% Number of Chanenls in EEG data
MICs=zeros(n,n);
for i=1:n
    for j=1:n
        one=Data(i,:);
        two=Data(j,:);
        miin=h(one')+h(two')-h([one',two']);
        MIC(i,j)=miin;
    end
end

I see different ways to estimate the MI's, however I had written my code above to verify.
When I tried to estimate the Mean(average) value of MI for all the pairs, Should I remove the diagonals values (which corresponds to the MI for its own pair) and the duplicates values (as the matrix is an outcome of the pairwise estimation).?
Or Should I just estimate the mean for the whole matrix which is an outcome of the above code?

4 个评论
显示 2更早的评论隐藏 2更早的评论

madhan ravi 2018-12-24

you didn't attach your data

Marimuthu Ananthavelu 2018-12-24

I have uploaded the data in CSV format due to its size. And Updated the code as well accordingly to load the data. Thanks for your willingness to help.

请先登录，再进行评论。

请先登录，再回答此问题。