ABSTRACT: The aim of this paper was to demonstrate possibilities of interrupted or inhomogeneous time series prediction. For this purpose, the stochastic Box-Jenkins models were used. The model was verified on monthly groundwater levels in the well Raková-západ Nr. 426 (27-years time series) and monthly yields of the spring Nižná Myšľa-Koscelek Nr. 2291 (28-years time series). The seasonal integrated mixed model SARIMA (1,0,0)x(3,l,1)12 was found to be an appropriate stochastic model for prediction of monthly groundwater levels of the well Raková-západ and SARIMA (l,l,0)x(l,l,0)12 for prediction of monthly yields of the spring Nižná Myšľa-Koscelek, in both cases with one year prediction time.
bibliografický odkaz na článok:
Némethyová, S. 1999: Prognózovanie porušených časových radov modelom SARIMA.- Podzemná voda ISSN 1335-1052, V, 2/1999, 54-63.
jazyk článku:
kľúčové slová:
prerušené a porušené časové rady, hladina podzemnej vody, výdatnosť prameňa, sezónny integrovaný model SARIMA
key words:
interrupted and inhomogeneous time series, groundwater level, spring yield, seasonal integrated mixed SARIMA model