ssfortran.SmootherResults#

class ssfortran.SmootherResults(filter, smoothed_state, smoothed_state_cov, scaled_smoothed_estimator, scaled_smoothed_estimator_cov, smoothed_measurement_disturbance, smoothed_measurement_disturbance_cov, smoothed_state_disturbance, smoothed_state_disturbance_cov)#

Output of the state and disturbance smoother (DK §4.4-4.5, §5.3).

Attributes:

filterFilterResults: The filter run the smoother used.
smoothed_statendarray, shape (m, n): \(\hat\alpha_t = E(\alpha_t | Y_n)\) (DK eq. 4.39).
smoothed_state_covndarray, shape (m, m, n): \(V_t = Var(\alpha_t | Y_n)\) (DK eq. 4.43).
scaled_smoothed_estimatorndarray, shape (m, n+1): \(r_t\) for t = 0, …, n.
scaled_smoothed_estimator_covndarray, shape (m, m, n+1): \(N_t\) for t = 0, …, n.
smoothed_measurement_disturbancendarray, shape (p, n): \(\hat\varepsilon_t = E(\varepsilon_t | Y_n)\). For a missing element this is its conditional mean given the observed elements, not zero.
smoothed_measurement_disturbance_covndarray, shape (p, p, n): \(Var(\varepsilon_t | Y_n)\).
smoothed_state_disturbancendarray, shape (r, n): \(\hat\eta_t = Q_t R_t' r_t\) (DK eq. 4.63).
smoothed_state_disturbance_covndarray, shape (r, r, n): \(Var(\eta_t | Y_n)\).