Moving Average

Moving Average Computation and Use

The software used for illustration purposes is Minitab, a well-known, easy-to-use, and reliable software [Page 634]package that runs on a PC. To start out, consider the following ordered data that were created to illustrate the calculation of a moving average (Figure 1).

The order refers simply to the fact that the data are in order. If the data are time series, then the order variable would be time at equally spaced intervals.

Because moving averages are often used to smooth data in the sense of eliminating nonmeaningful irregularities in the data, the best practice is to first plot the data to see what they look like. The questions that a plot can answer are, Are there any irregularities in the data that are not meaningful? Do the data display some sort of trend that might be meaningful?

The plot in Figure 2 was produced in Minitab by simply selecting Graph from the main menu and then Time Series Plot. Notice that the data have a slight upward trend with one valley and a couple of peaks.

To explore a moving average, consider a five-term moving average; the number of terms in the moving average is usually determined by trial and error if no other insight is available.

The hand calculations for the first term of the moving average are to simply average the first five data points:

where MA(1) is the first smoothed value and data(n) is the data column value at Period or Order 1.

The second smoothed value is

Thus, the calculations continue by shifting down one step and computing the average of the next five terms—thus the term moving average.

...