Data Analysis with Microsoft Excel: Updated for Office 2007

Chapter 11 Times Series 461

With a higher smoothing constant, the forecasted trend of the time series

shows an increase of 0.044 points per unit of time. The area curve indicates

that the smoothed trend estimate has a shorter memory; only the most re-

cent observations are relevant in estimating the trend.

Using this worksheet, you can change the values of the smoothing con-

stant for the location and trend parameters. What combinations result in the

lowest values for the standard error? When you are fi nished with your inves-

tigations, close the workbook. You do not have to save any of your changes.

Now let’s return to the global temperature data. In using one-parameter

exponential smoothing the forecasted values underestimated the most recent

trend in temperature data. To compensate we’ll use two-parameter exponential

smoothing in an attempt to “pick up” the most recent trend of increasing

temperatures.

To forecast global temperatures using two-parameter exponential

smoothing:

1 Return to the Global Temperature Analysis workbook and go to the

Temperature worksheet.

2 Click Time Series from the StatPlus menu and then click Exponen-

tial Smoothing.

3 Click the Data Values button and select Fahrenheit from the list of

range names. Click OK.

Figure 11-19

Increasing
the value of
t to 0.40 a negative trend
is forecast

only the most

recent observations

are used in

forecasting the trend