The term hidden state is used when analysing a time series where some of the important aspects of the phenomenon being observed are not in the measured time series. For example, when observing the weather, we may know windspeeds and temepratures at ground level, but not at higher altitudes. In order to make predictions, this hidden state has to be modelled or inferred based on the observable values.
Used in Chap. 14: pages 205, 206, 207, 210, 212, 219