Receives the Kalman deviance (−2·logL, scalar np.Array), the current
parameter vector in natural scale (np.Array), and a layout
descriptor (DlmParamMeta). Must return a scalar np.Array.
All operations on the inputs must be AD-safe (jax-js ops only).
params layout: [s?, w₀, …, w_{m-1}, φ₁, …, φ_p]
s and wᵢ are positive std devs (not log-transformed).
Custom loss function type for MAP estimation.
Receives the Kalman deviance (−2·logL, scalar
np.Array), the current parameter vector in natural scale (np.Array), and a layout descriptor (DlmParamMeta). Must return a scalarnp.Array. All operations on the inputs must be AD-safe (jax-js ops only).paramslayout:[s?, w₀, …, w_{m-1}, φ₁, …, φ_p]sandwᵢare positive std devs (not log-transformed).φⱼare unconstrained AR coefficients.