Time Series Forecasting Using QM for Windows

[Page 690 ( continued )]

QM for Windows has the capability to perform forecasting for all the time series methods we have described so far. QM for Windows has modules for moving averages, exponential smoothing and adjusted exponential smoothing, and linear regression.

To demonstrate the forecasting capability of QM for Windows, we will generate the exponential smoothing ( a = .30) forecast computed manually for PM Computer Services (Table 15.4). The solution output is shown in Exhibit 15.6.

Exhibit 15.6.

(This item is displayed on page 691 in the print version)

Notice that the solution summary includes the forecast per period and the forecast for the next period (13), as well as three measures of forecast accuracy: average error (bias), mean absolute deviation ( MAD ), and mean squared error ( MSE ).

---

[Page 691]

The least squares module or the simple linear regression module in QM for Windows can be used to develop a linear trend line forecast. Using the least squares module, the solution summary for the linear trend line forecast we developed for PM Computer Services is shown in Exhibit 15.7.

Exhibit 15.7.