Chapter 4: Noncausal Gaussian Models
Charles A. Bouman · Society for Industrial and Applied Mathematics eBooks · 2022
One disadvantage of AR processes is that their construction depends on a causal ordering of points in time. For many applications, it is completely reasonable to order points into the future, present, and past. For example, in real-time processing of audio signals, this is a very natural organization of the data. But sometimes measurements have no natural ordering. For example, the temperatures measured along a road are a 1-D signal, but the direction of causality is not well defined. Which end of the road should represent the past and which the future?