statsmodels.tsa.statespace.kalman_smoother.SmootherResults¶

class statsmodels.tsa.statespace.kalman_smoother.SmootherResults(model)[source]¶

Results from applying the Kalman smoother and/or filter to a state space model.

Parameters:¶

model : Representation¶

A Statespace representation

nobs¶

Number of observations.

Type:¶

int

k_endog¶

The dimension of the observation series.

Type:¶

int

k_states¶

The dimension of the unobserved state process.

Type:¶

int

k_posdef¶

The dimension of a guaranteed positive definite covariance matrix describing the shocks in the measurement equation.

Type:¶

int

dtype¶

Datatype of representation matrices

Type:¶

dtype

prefix¶

BLAS prefix of representation matrices

Type:¶

str

shapes¶

A dictionary recording the shapes of each of the representation matrices as tuples.

Type:¶

dictionary of name:tuple

endog¶

The observation vector.

Type:¶

ndarray

design¶

The design matrix, \(Z\).

Type:¶

ndarray

obs_intercept¶

The intercept for the observation equation, \(d\).

Type:¶

ndarray

obs_cov¶

The covariance matrix for the observation equation \(H\).

Type:¶

ndarray

transition¶

The transition matrix, \(T\).

Type:¶

ndarray

state_intercept¶

The intercept for the transition equation, \(c\).

Type:¶

ndarray

selection¶

The selection matrix, \(R\).

Type:¶

ndarray

state_cov¶

The covariance matrix for the state equation \(Q\).

Type:¶

ndarray

missing¶

An array of the same size as endog, filled with boolean values that are True if the corresponding entry in endog is NaN and False otherwise.

Type:¶

array of bool

nmissing¶

An array of size nobs, where the ith entry is the number (between 0 and k_endog) of NaNs in the ith row of the endog array.

Type:¶

array of int

time_invariant¶

Whether or not the representation matrices are time-invariant

Type:¶

bool

initialization¶

Kalman filter initialization method.

Type:¶

str

initial_state¶

The state vector used to initialize the Kalamn filter.

Type:¶

array_like

initial_state_cov¶

The state covariance matrix used to initialize the Kalamn filter.

Type:¶

array_like

filter_method¶

Bitmask representing the Kalman filtering method

Type:¶

int

inversion_method¶

Bitmask representing the method used to invert the forecast error covariance matrix.

Type:¶

int

stability_method¶

Bitmask representing the methods used to promote numerical stability in the Kalman filter recursions.

Type:¶

int

conserve_memory¶

Bitmask representing the selected memory conservation method.

Type:¶

int

tolerance¶

The tolerance at which the Kalman filter determines convergence to steady-state.

Type:¶

float

loglikelihood_burn¶

The number of initial periods during which the loglikelihood is not recorded.

Type:¶

int

converged¶

Whether or not the Kalman filter converged.

Type:¶

bool

period_converged¶

The time period in which the Kalman filter converged.

Type:¶

int

filtered_state¶

The filtered state vector at each time period.

Type:¶

ndarray

filtered_state_cov¶

The filtered state covariance matrix at each time period.

Type:¶

ndarray

predicted_state¶

The predicted state vector at each time period.

Type:¶

ndarray

predicted_state_cov¶

The predicted state covariance matrix at each time period.

Type:¶

ndarray

kalman_gain¶

The Kalman gain at each time period.

Type:¶

ndarray

forecasts¶

The one-step-ahead forecasts of observations at each time period.

Type:¶

ndarray

forecasts_error¶

The forecast errors at each time period.

Type:¶

ndarray

forecasts_error_cov¶

The forecast error covariance matrices at each time period.

Type:¶

ndarray

loglikelihood¶

The loglikelihood values at each time period.

Type:¶

ndarray

collapsed_forecasts¶

If filtering using collapsed observations, stores the one-step-ahead forecasts of collapsed observations at each time period.

Type:¶

ndarray

collapsed_forecasts_error¶

If filtering using collapsed observations, stores the one-step-ahead forecast errors of collapsed observations at each time period.

Type:¶

ndarray

collapsed_forecasts_error_cov¶

If filtering using collapsed observations, stores the one-step-ahead forecast error covariance matrices of collapsed observations at each time period.

Type:¶

ndarray

standardized_forecast_error¶

The standardized forecast errors

Type:¶

ndarray

smoother_output¶

Bitmask representing the generated Kalman smoothing output

Type:¶

int

scaled_smoothed_estimator¶

The scaled smoothed estimator at each time period.

Type:¶

ndarray

scaled_smoothed_estimator_cov¶

The scaled smoothed estimator covariance matrices at each time period.

Type:¶

ndarray

smoothing_error¶

The smoothing error covariance matrices at each time period.

Type:¶

ndarray

smoothed_state¶

The smoothed state at each time period.

Type:¶

ndarray

smoothed_state_cov¶

The smoothed state covariance matrices at each time period.

Type:¶

ndarray

smoothed_state_autocov¶

The smoothed state lago-one autocovariance matrices at each time period: \(Cov(\alpha_{t+1}, \alpha_t)\).

Type:¶

ndarray

smoothed_measurement_disturbance¶

The smoothed measurement at each time period.

Type:¶

ndarray

smoothed_state_disturbance¶

The smoothed state at each time period.

Type:¶

ndarray

smoothed_measurement_disturbance_cov¶

The smoothed measurement disturbance covariance matrices at each time period.

Type:¶

ndarray

smoothed_state_disturbance_cov¶

The smoothed state disturbance covariance matrices at each time period.

Type:¶

ndarray

Methods

get_smoothed_decomposition([...])	Decompose smoothed output into contributions from observations
news(previous[, t, start, end, ...])	Compute the news and impacts associated with a data release
predict([start, end, dynamic])	In-sample and out-of-sample prediction for state space models generally
smoothed_state_autocovariance([lag, t, ...])	Compute state vector autocovariances, conditional on the full dataset
smoothed_state_gain(updates_ix[, t, start, ...])	Cov(tilde alpha_{t}, I) Var(I, I)^{-1}
update_filter(kalman_filter)	Update the filter results
update_representation(model[, only_options])	Update the results to match a given model
update_smoother(smoother)	Update the smoother results

Properties

kalman_gain	Kalman gain matrices
smoothed_forecasts
smoothed_forecasts_error
smoothed_forecasts_error_cov
standardized_forecasts_error	Standardized forecast errors