ABSTRACT
In this chapter we bring together the different approaches presented in earlier chapters. First, we see how partial coherency graphs can be applied in the frequency domain to extend insight into the graphical modeling approach of Chapter 5. Second, we show how spectral analysis can be used directly to estimate the impulse response relationships between time series even in the context of feedback between the series. We also extend this approach to continuous, or high frequency sampled, time series by applying the frequency warp associated with VCZAR models. Finally, we show how partial correlation graphs can be extended to identify structural forms of the VZAR model.
