Asymptotic Behavior of Time Series Aggregates.
WISCONSIN UNIV MADISON DEPT OF STATISTICS
Pagination or Media Count:
The paper discusses the efficiency of disaggregation in forecasting time series aggregates. Let be the disaggregated series and XTZMT-m1 ... ZMT be the m-component aggregated series. Forecasts of future XT may be constructed from data on i Zt or ii XT. It is shown that, for large m, there is no gain in using the disaggreagated data if Zt is stationary, but dramatic gain can be obtained when Zt is non-stationary. Author
- Statistics and Probability