The fisher submodule#

This submodule implements the Fisher information matrix for cMHN and oMHN.

It contains function for computing the FIM in Cython and for calling the CUDA implementation.

mhn.full_state_space.fisher.cython_fisher(double[:, :] theta)#

Computes the fisher information matrix for cMHN. The formulas are described in S. Vocht. Identifiability of Mutual Hazard Networks. Unpublished bachelor thesis, 2022

Parameters:: theta (np.ndarray) – matrix containing the theta values
Returns:: The Fisher information matrix for theta
Return type:: np.ndarray

mhn.full_state_space.fisher.fisher(ndarray log_theta: np.ndarray, omhn: bool = True, use_cuda: bool = False) → np.ndarray#

Computes the Fisher information matrix for o/cMHN.

Parameters:

log_theta (np.ndarray) – matrix containing the log theta values
omhn (bool, optional) – whether MHN is oMHN or cMHN
use_cuda (bool, optional) – whether to use GPU acceleration

Returns:

The Fisher information matrix for log_theta

Return type:

np.ndarray

mhn.full_state_space.fisher.omhn_fisher(ndarray log_theta: np.ndarray, use_cuda: bool = False) → np.ndarray#

Computes the Fisher information matrix for oMHN.