Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Cited by Google
- Similars in SciELO
- Similars in Google
Share
Revista Colombiana de Estadística
Print version ISSN 0120-1751
Abstract
CASTANO, ELKIN. Time Series Data Reconstruction: An Application to the Hourly Demand of Electricity. Rev.Colomb.Estad. [online]. 2007, vol.30, n.2, pp.247-263. ISSN 0120-1751.
Usually, in the identification and estimation of ARIMA models it is supposed that the series to analyze contain neither missing data, nor atypical observations, and interventions do not exist under study period. Nevertheless, in the practice, these problems can happen simultaneously, affecting the identification of the suitable model and therefore his forecasting capacity. This article presents a procedure that allows to estimate the effect of the interventions, of the atypical observations, to estimate the missing observations and simultaneously to identify the ARIMA model. The procedure is applied to a series of hourly electricity demand in which the three mentioned events happen.
Keywords : Atypical observations; Missing observations; Intervention; Transfer function; ARIMA.