p values from correlation matrix using nancov

Question

JacquelineK 2017-9-21

评论： JacquelineK 2017-9-23

I have a matrix A 225x13 containing missing values.

Because of the missing values I can not use corrcoef for my correlation matrix. Nor can I use [r,p].

So my current code looks like this:

Cov_A = nancov(A)
R = corrcov(Cov_A)

With corrcoef it would definetely be easier to just:

[r,p] = corrcoef(B);
[i,p] = find(p<0.05);
[i,p]

How do I get the same results, meaning r and p with "corrcov"? Or is it even possible to get pvalues with corrcov??

Answer 1

Brendan Hamm 2017-9-22

The corrcoef function accepts Name-Value pair inputs to handle missing data.

[r,p] = corrcoeff(B,'rows','complete')
[r,p] = corcoeff(B,'rows','pairwise')

JacquelineK 2017-9-23

THANKS :)