MarkovSwitchingARResult._summary_table#

MarkovSwitchingARResult._summary_table()[source]#

Structured summary rendered by every display path.

Metadata carries the specification, regime count, EM iteration count and convergence flag, the sample size, the likelihood with its criteria, and the posterior entropy; the coefficient table is params in order; the notes lead with a non-convergence warning when it applies, then the labelling convention, the separation and durations, the stationarity caveat, the redundancy of the transition rows, and the untestability of K.

Return type:

SummaryTable

Returns:

The SummaryTable.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> model = MarkovSwitchingAR(rng.standard_normal(300), order=1)
>>> table = model.fit(seed=0, n_init=2)._summary_table()
>>> table.title, len(table.notes), table.metadata[6]
('MSIH(2)-AR(1) Results', 6, ('Converged', 'True'))
>>> model.fit(seed=0, n_init=2, max_iter=1)._summary_table().notes[0][:21]
'EM did NOT converge i'