DETERMINISTIC PROCESS

(Statistics) An analytical and forecasting technique which assumes that the future can be predicted exactly from its past. Consequently, it is assumed that the data series to be forecasted contains all the information necessary to predict its future behavior. A deterministic process or relationship is assumed to be "exact", and therefore assumes no presence of a Disturbance (or Error) Term. The simplest form of this process is commonly termed an AutoRegressive Moving Average (ARMA) Process, or Box-Jenkins, which involves regressing a series on itself and using solely the historical patterns contained in the data to formulate forecasts. As a naive method, such a process does not include the capability to incorporate external "shocks" or other influences which may have an effect on the future behavior of a series. Such a technique is typically used only for well-behaved data showing typically predictable repetitive cycles and patterns. Contrast with Stochastic Process.