ARMA MODELS OF THE SIGNAL, SPECTRUM AND AUTOCORRELATION FUNCTION OF THE RIVERS AMBIENT NOISE

Miodrag Vračar

Military Technical Institute, Belgrade, vracarmiodrag@open.telekom.rs

Abstract: The dependences are very common in time series observations. One of the approaches of modeling time series dependence is ARMA model, acronym of Autoregressive Moving Average. ARMA model provides one of the basic tools in time series modeling. Time series is possible to model on a way that the level of its current observations depends on the level of its lagged observations. In this paper, such approach is confirmed on three examples of the river ambient noise time series that were measured in the rivers the Danube, the Sava and the Tisa. The quality of ARMA time series estimates is verified by calculation spectra and autocorrelation function.