Quantitative and Qualitative Forecasting

.l-uriRc,rsrrxr; clnpter I B 461

Straight Line Graph of Deseasonalized Equation

5.000

.+.500

,+,000

3,500

3,000

Sales 2.-500 2.000

1 ,i00 1.000

500

0

D

w -#* i:or the Excel template. visit

\M.mhhe.com,'jacobs 1 4e.

561 Quarters

ll

step 5. create the final forecast by adjusting the regression line by the seasonal factor. Recall that the Iequation has been deseasonalized. We now reverse the procedure by multiplying the quarterly dara we derived by the seasonal factor for that quarrer:

Ymorr Fonrc.r,sr Prnroo Qr,;rnrsx REGREssToN LrNE SersoNrl Frcrop (y X SE,\soNAL FACroRl

1.2i0

13 1 142 15 3 164

5,003.5 5,345.7 5,687.9 6,030.'i

0.82 1.10 0.97 1.12

4J42.87 5,884.27 5,517.26 6,753.71

Our lbrecast'is now complete.

Wlren a straight line is fitted through data points and then usecl tbr lbrccasting, enlxs can come from two sourccs. First, there are the usual errors similar to the standard clevia- tion of any sct of data. Sccond, thcrc are errors that arisc becansc the linc is wrong. E,xhibit 18. l-3 shou's this en'or range. Instead of developing thc statistics here. we will briefly show why thc range broadens. First. visualize that one line is drau,n that has somc crror such that it slants too steeply upward. Standard errors are then calculated fbr this line. Now vi- sualizc another line that slants too stceply downward. It also has a standard crrclr. Thc total

Prediction Intervals for Linear Trend

.,1' ,"Prediction

u. inte-rval

Prediction interval --.-'@

'.,:.t '-."i:

Past

Tiine

exhibit ltt.l

exhibit 18.1

Present Future

Demand