How to omit certain correlations using CORR on 2 matrices?

Hi Everyone,

I have two matrices with the size of 40 (observations) x 68 (variables). I have used corr to plot the correlation matrix between both datasets, yielding a 68 x 68 correlation matrix.

How may I omit correlations between measurements in corresponding variables, or turn the values into NaN? After this, I am using FDR to correct the p values.

Thanks.

[Brian Hannan]

Hi Shane, in order for us to help you with this, can you tell me how you are deciding which correlation values you want to throw out?

[ShaneS]

Hi,

I don't want to look at correlations in the same positions across both datasets.

For example, no correlations should be done between Dataset1(1,1) and Dataset_B(1,1), etc.

Thanks.