Zero correlation coefficient, example

This question of mine is not tightly related to Matlab, but is relevant to it:

I'm looking how to fill in this matrix in a few nontrivial ways

[[a,b,c],[d,e,f]] so that as many places in the ans of this matrix

corrcoef([a,b,c],[d,e,f])

are zero. My attempts yield NaN result in most cases.

[Jan]

I'm still unable to obtain 0 in ans in my corrcoef matrix that avoids NaN.

The best I was able is to do this:

corrcoef([1,2,3];[-1,2,1])

Can anyone provide me with any example that creates 0 instead of NaN ?

I have both, the difficulty to create 0 in my Ans and understand the reasons provided in @Adam Danz Answer.

[Adam Danz]

That line throws an error. There are two ways to correct it but you need to let us know which method fits your needs. See the corrcoef documentation to learn the difference between the two syntaxes.

Original line

corrcoef([1,2,3];[-1,2,1])

Invalid expression. When calling a function or indexing a variable, use parentheses. Otherwise, check for mismatched delimiters.

Error in connector.internal.fevalMatlab

Error in connector.internal.fevalJSON

Corrected version 1

corrcoef([1,2,3],[-1,2,1])

ans =

1.0000 0.6547

0.6547 1.0000

Corrected version 2

corrcoef([1,2,3 ; -1,2,1])

ans =

1.0000 NaN 1.0000

NaN NaN NaN

1.0000 NaN 1.0000

[Jan]

I need the second one.

[Adam Danz]

x = [1,2,3 ; -1,2,1]

x = 2×3

1 2 3 -1 2 1

corrcoef computes the correlation coeficient for each pair of columns. Let's look at column 1 and 2

x(:,1), x(:,2)

ans = 2×1

1 -1

ans = 2×1

2 2

corr(x(:,1),x(:,2))

ans = NaN

Notice that x(:,2) contains the same values. Now look at the link in my answer and pay attantion to the part that says, "When each value of y is identical, the result is a vector of 0s. When you divide by zero, you get NaN. ".

Answer 1

[Adam Danz]

This answer explains why NaN values may appear in correlations.

Make sure your inputs avoid those errors. If you have trouble understanding the reasons explained in that answer or if you have trouble creating inputs that avoid NaNs after considering those reasons, share your inputs and we can help guide you in the right direction.