SVResult.state_space

property SVResult.state_space: _NonlinearStateSpace

The fitted model on the nonlinear substrate.

Carries the multiplicative measurement through the observation_loglik hook, so only the particle filter reads it; particle_filter on the estimation sample estimates the exact likelihood, and on new data reads that data with this fit.

Example

A particle estimate of the exact likelihood at the QML point:

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> y = rng.standard_normal(200)
>>> res = SV(y, mean="zero").fit()
>>> ssm = res.state_space
>>> ssm.k_states, ssm.k_endog
(1, 1)
>>> bool(np.isfinite(ssm.particle_filter(y, n_particles=200, seed=0).loglikelihood))
True