In the previous chapter, you learned how to identify and forecast a random walk process. We defined a random walk process as a series whose first difference is stationary with no autocorrelation. This means that plotting its ACF will show no significant coefficients after lag 0. However, it is possible that a stationary process may still exhibit autocorrelation. In this case, we have a time series that can be approximated by a moving average model MA(q)), an autoregressive model AR(p)), or an autoregressive moving average model ARMA(p,q). In this chapter, we will focus on identifying and modeling using the moving average model.
Suppose that you want to forecast the volume of widget sales from the XYZ Widget Company. By predicting futures sales, the company will be able to better manage its production of widgets and avoid producing too many or too few. If not enough widgets are produced, the company will not be able to meet their clients’ demands, leaving customers unhappy. On the other hand, producing too many widgets will increase inventory. The widgets might become obsolete or lose their value, which will increase the business’s liabilities, ultimately making shareholders unhappy.