17 State space models and the Kalman filter
An ARMA model describes a series directly. A state space model supposes, instead, an unobserved process driving it, and the observations are a noisy view of that process. The framework is more general (every ARMA model can be written in state space form) and it handles missing data and time-varying parameters, which ARMA cannot.
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.