The data likelihood, given a fully transformed latent state, eta, and an observation the log-likelihood can be calculated for use in inference algorithms
The Linear, deterministic transformation function.
The Linear, deterministic transformation function. f is used to add seasonal factors or other time depending linear transformations
The observation model, a function from eta to a distribution over the observations realisations can be produced from the observation model by calling draw
An exact or approximate solution to a diffusion process, used to advance the latent state.
An exact or approximate solution to a diffusion process, used to advance the latent state. This function returns a distribution over the next state and can be simulated from
The linking-function, transforms the state space into the parameter space of the observation distribution using a possibly non-linear transformation