schur_information¶
- ell1fit.information.schur_information(joint, n_shared, name)[source]¶
Split one segment’s joint information matrix into raw and effective parts.
- Parameters:
- jointarray-like
The observed information (minus the log-likelihood Hessian) over the shared parameters first, then that segment’s own spin parameters.
- n_sharedint
How many leading rows and columns are the shared parameters.
- namestr
Identifier for the segment, carried through into errors and tables.
- Returns:
- SegmentInformation
- Raises:
- numpy.linalg.LinAlgError
If the segment’s own spin block is singular, i.e. the data cannot pin down its own spin parameters. The message names the segment.