ABSTRACT
The Auto Regressive Integrated Moving Average (ARIMA) approach is a set of powerful statistical techniques through which the exact nature of the serial dependency of a set of observations made over a period of time can be assessed. The ARIMA approach consists of a large array of highly complex mathematical techniques for time series analysis. The development of ARIMA techniques is largely due to the work of Box and Jenkins. The objective of the ARIMA analysis is to describe the nature of the serial dependency of the time series in exact mathematical terms. ARIMA requires a fundamental set of statistical principles from which statistical techniques can be derived. The time series model departs from the Classical Theory of measurement model when the stochastic component is considered. A nonseasonal stochastic process can be described by various combinations of three basic characteristics known as model parameters: autoregressive (AR) function, a differencing (I) function, and a moving average (MA) function.
