fips.linalg.KroneckerOperator.logdet#

KroneckerOperator.logdet()[source]#

Return the log-determinant through the factors.

When the rows are the whole product, in any order, \(\log\det S = 2 \sum_i \log s_i + \sum_k (N / N_k) \log\det C_k\), where \(s\) are the standard deviations, \(N_k\) the size of factor k and \(N\) their product. For a strict subset of the product the dense matrix is factored.

Returns:

log(det(S)).

Return type:

float

Raises:

numpy.linalg.LinAlgError – If a factor is not positive-definite or a standard deviation is zero.