Correlation between multiples items

Hello, Could I get a range of items less correlated from a correlation matrix?

For example from:

A, B, C, D, E ... N

A,

B

C

D

E

...

N

Output: items less correlated are ''A,D,E''

I only need a multiple combination of items that are less correlated each others. (A - D), (A - E), (D - E)

Sorry for my English...

Thank you everyone!

Best regards

[Subhadeep Koley]

Hi Davide, I assume you already have the correlation matrix and you need only some items which are least correlated. Am I right ?

[Davide Sinesi]

Yes

[John D'Errico]

https://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer

Seriously, this question has nothing to do with MATLAB. Your question is too vague to have an answer.

So spend some time and think about exactly what you need, in terms of mathematics (or statistics). That may involve asking questions on another forum, perhaps a statistics forum. Until you define what problem you are trying to solve, you cannot formulate it in terms of code, in any language.

[John D'Errico]

By the way, your question, if I interpret it as I wonder if I know where you may be going, may have serious problems anyway.

For example, suppose that you see that A is not highly correlated to B? As well, you might see that B is not highly correlated to C. However, it can still be true that A and C are highly correlated. In fact, A and C may be perfectly correlated to each other. Consider the matrix:

C = [1 0 1;

0 1 0;

1 0 1]

So we have A and B with a zero correlation. Likewise, B and C have zero correlation. But A and C are perfectly correlated.

The relation, "not highly correlated" is not a transitive relation. So be careful in any conclusions that you may draw.

Answer 1

[Subhadeep Koley]

Use the function below to find out n least correlated items,

function findLeastCorr(corr)
% FINDLEASTCORR finds least correlated items from a given correlation matrix
% INPUT
% -----
% corr_cell - User's correlation matrix
% 
% EXAMPLE
% -------
% findLeastCorr(yourCorrMatrix);
% 
% 
sz = size(corr);
% Sorting from lowest correlatin to highest
[~, corr_ind] = sort(corr(:));
% Finding the index of the least correlated items
prompt = 'How many least correlated items you want?';
n = input(prompt);
items = {1,1,n};
for i=1:n
    [a, b] = ind2sub(sz,corr_ind(i));
    items{i} = [a, b]; 
end
% Identify unique elements
items_mat = cell2mat(items);
items_unique = unique(items_mat);
% Displaying the items
disp(['Items less correlated are: ',num2str(items_unique(1:end))]);
end

Hope this helps!