CL_cov2cor

Description

• Computes the correlation matrix and the standard deviation vector from the covariance matrix.

  The correlation matrix and standard deviation vector are defined as follows:

  - corm(i,i) = 1

  - corm(i,j) = covm(i,j) / (sd(i) * sd(j))

  - sd(i) = covm(i,i)^(1/2)

  Notes:

  - The covariance matrix should be symetrical (not checked).

  - An error is raised if one diagonal term is not positive.

  - If one diagonal term of the covariance matrix is 0, the corresponding diagonal term of the correlation matrix is 1; other terms of the correlation matrix on the same row or column are set to %nan.

Parameters

covm:

Covariance matrix (NxNxK)

corm:

Correlation matrix (NxNxK)

sd:

Standard deviation vector (NxK)

Authors

• CNES - DCT/SB

See also

• CL_cor2cov

Examples

covm = [1, 0, 1.5; 0, 4, -4.8; 1.5, -4.8, 9]; [corm, sd] = CL_cov2cor(covm); // Consistency check: covm2 = CL_cor2cov(corm, sd) // => covm