View source on GitHub
|
TensorFlow Probability experimental sequential estimation package.
Classes
class EnsembleKalmanFilterState: State for Ensemble Kalman Filter algorithms.
class IteratedFilter: A model augmented with parameter perturbations for iterated filtering.
Functions
ensemble_adjustment_kalman_filter_update(...): Ensemble Adjustment Kalman Filter Update.
ensemble_kalman_filter_log_marginal_likelihood(...): Ensemble Kalman filter log marginal likelihood.
ensemble_kalman_filter_predict(...): Ensemble Kalman filter prediction step.
ensemble_kalman_filter_update(...): Ensemble Kalman filter update step.
extended_kalman_filter(...): Applies an Extended Kalman Filter to observed data.
geometric_cooling_schedule(...): Defines a cooling schedule following a geometric sequence.
inflate_by_scaled_identity_fn(...): Return function that scales up covariance matrix by scaling_factor**2.