Temporal Disaggregation of Time Series Revisited

Abstract

With the benefits of increased computing power and much improved software, temporal disaggregation is examined. Disaggregation, the process of obtaining high frequency data from low frequency data has been discussed for many years. This study examines three methods which utilize the autoregressive integrated moving average (ARIMA) model in a simulation study comparing parameter estimation, disaggregation mean square error, and forecast mean square error. Finally, the three methods are applied to a real-world time series.