Time series forecasts based on transfer function theory
Larry Duane Gianchetta · OakTrust (Texas A&M University Libraries) · 1975
This dissertation is concerned with recent techniques presented by Box and Jenkins [1970] dealing with transfer function forecast models. The first consideration is what happens to the forecasts and their variances, as presented by Box and Jenkins [1970], given the special case of knowing the value of the leading indicator variable one lead time into the future and forecasting the variable of interest for the corresponding time period. The second consideration deals with the shortcomings of the Box and Jenkins [1970] techniques of estimation and identification given certain probably situations and alternative methods for estimation and ident cation. Finally, we present an example incorporating both of the aforementioned considerations. ...