NonlinearSSM.extended_filter

NonlinearSSM.extended_filter(y)

The extended Kalman filter: linearize, then filter exactly.

First-order accurate in the nonlinearity; exact when f and g are linear, in which case it is the Kalman filter. Jacobians come from central differences, so the maps need to be smooth at the working point but need not expose derivatives.

Parameters:

y (ArrayLike) – Data, (n, p); numpy.nan rows are missing.

Return type:

_KalmanFilterResult

Returns:

The standard filter record; the log-likelihood is the linearized approximation.

Raises:

• SpecificationError – If the model carries custom hooks.

• NumericalError – If the filter diverges.