ABSTRACT
We introduced the multivariate autoregressive or VAR(p) model (2.7) in Chapter 2. It is the most widely used practical tool for modeling multivariate time series and in this chapter we explain and illustrate how it can be fitted to observed series by estimating the model parameters. We consider three methods of estimation:
1. solving the Yule-Walker equations presented in (2.5), but replacing the autocovariances Γi,j,k with the sample autocovariances Ci,j,k defined in (3.35).
