The preceding two sections describe the auto-regressive model AR(p), which regresses on its own lagged terms and moving average model MA(q) builds a function of error terms of the past. The AR(p) models tend to capture the mean reversion effect whereas MA(q) models tend to capture the shock effect in error ,which are not normal or unpredicted events. Thus, the ARMA model combines the power of AR and MA components together. An ARMA(p, q) time series forecasting model incorporates the pth order AR and qth order MA model, respectively.
The ARMA (1, 1) model is represented as follows:
The ARMA(1, 2) model is denoted ...
