Filters

Forward filtering algorithms for state-space models. Dynaris provides four filtering algorithms: the Kalman filter for linear-Gaussian models, and the Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), and Particle Filter for nonlinear models.

Kalman Filter

Forward filtering for linear-Gaussian state-space models. The Kalman filter processes observations sequentially, computing the posterior state distribution at each time step.

Note

Most users do not need to call these functions directly — DLM.fit() wraps the Kalman filter internally. These are available for advanced use cases requiring direct access to intermediate filter quantities.

Extended Kalman Filter (EKF)

Linearizes nonlinear transition and observation functions at each time step using automatic Jacobians via jax.jacfwd, then applies the standard Kalman predict/update equations to the linearized system.

Unscented Kalman Filter (UKF)

Propagates sigma points through nonlinear transition and observation functions to capture the posterior mean and covariance without linearization. Uses the scaled unscented transform with configurable alpha, beta, kappa parameters.

Particle Filter (SMC)

Bootstrap particle filter / Sequential Monte Carlo. Represents the filtering posterior as a weighted set of particles, enabling inference in nonlinear, non-Gaussian state-space models. Supports multinomial, systematic, and stratified resampling strategies.