statsmodels.tsa.statespace.kalman_filter.KalmanFilter.initialize#
- KalmanFilter.initialize(initialization, approximate_diffuse_variance=None, constant=None, stationary_cov=None, a=None, Pstar=None, Pinf=None, A=None, R0=None, Q0=None)#
Create an Initialization object if necessary
- Parameters:
- initialization
strorInitialization Initialization method for the initial state. If a string, must be one of {‘known’, ‘components’, ‘approximate_diffuse’, ‘stationary’, ‘diffuse’}. Otherwise, may be an already-created Initialization object, which is used directly.
- approximate_diffuse_variance
float,optional Initial variance used when initialization=’approximate_diffuse’ is specified. Default is self.initial_variance.
- constantarray_like,
optional Known mean of the initial state vector, used when initialization=’known’.
- stationary_covarray_like,
optional Known covariance matrix of the initial state vector, used when initialization=’known’.
- aarray_like,
optional Vector of constant values describing the mean of the stationary component of the initial state, used when initialization=’components’.
- Pstararray_like,
optional Stationary component of the initial state covariance matrix, used when initialization=’components’.
- Pinfarray_like,
optional Diffuse component of the initial state covariance matrix, used when initialization=’components’.
- Aarray_like,
optional Diffuse selection matrix, used when initialization=’components’.
- R0array_like,
optional Stationary selection matrix, used when initialization=’components’.
- Q0array_like,
optional Covariance matrix associated with stationary initial states, used when initialization=’components’.
- initialization