Box-Jenkins method is frequently used for time series analysis and forecasting, by using the so-called ARIMA family of models. A key element in the Box-Jenkins procedure is the so-called PACF which stands for Partial Autocorrelation Function, see for instance, Bisgaard and Kulahci [1]. Box et al. [2] gave the general formula of the PACF for the first-order Moving Average Model (MA(1)), which is a member of the family of ARIMA models. The main objective of this paper is to present a simple and an alternative derivation of the PACF of the MA(1) model.

Box-Jenkins procedure, ARIMA models, time series analysis.