bayespy.utils.random.covariance

bayespy.utils.random.covariance(D, size=(), nu=None)[source]

Draw a random covariance matrix.

Draws from inverse-Wishart distribution. The distribution of each element is independent of the dimensionality of the matrix.

C ~ Inv-W(I, D)

Parameters:
  • D (int) – Dimensionality of the covariance matrix.

  • Returns

  • --------

  • C ((D,D) ndarray) – Positive-definite symmetric D\times D matrix.