The sum of the stochastic pattern and random errors when a stochastic process is assumed to be the data-generating process. The distinction between process and realization is relevant when considering, for example, the difference between theoretical and sample autocorrelation and between theoretical and sample partial autocorrelation functions.