ABSTRACT
Multivariate time series analysis develops models and methodology for simultaneous description and forecasting of multiple time series that have a common underlying structure. This chapter shows how univariate analyses of such time series can be useful in discovering some of the common latent structure that may be underlying the multiple components of a time series vector. The results obtained from these univariate models can also be enlightening in terms of motivating some of the multivariate models that will be discussed later. We also introduce some simple multivariate models that are particular cases of much more sophisticated and structurally rich models that will be discussed in subsequent chapters.
