Prediction of multivariate time series by autoregressive model fitting (Q1067337)

In this paper the authors study the prediction errors in time series when the predictors themselves are estimated. In particular, stationary vector series are studied, however, without attention to their structure. In other words, the processes are tacitly assumed to be ''generic''. The main results found earlier for scalar processes, such as the mean square prediction error, are generalized to the studied vector case. I would like to add that to my knowledge the first such basic result in the scalar case was found by \textit{L. D. Davisson}, IEEE Trans. Inf. Theory IT-11, 527-532 (1965; Zbl 0152.175), which the authors appear to be unaware of.