Deterministic method for the prediction of time series

Abstract

The business analyst is frequently forced to make decision using data on a certain business process obtained within a short time interval. Under these conditions, the analyst is not in a position to use traditional statistical methods and should be satisfied with experimental samples that are few in number. The paper deals with the new method for the prediction of time series based on the concepts of system identification. The suggested method shows a reasonable flexibility and accuracy when the analyzed process possesses a certain regularity property, and it is usable for the prediction of time series in a short-term perspective.