Microsoft Office Tutorials and References
In Depth Information
Exhibit 3.23 Summary output for product E quarter 1
measures how well the estimated values of the regression line correspond to the
actual quarterly sales data; it is a guide to the goodness of fit of the regres-
sion model. R-square values can vary from 0 to 1, with 1 indicating perfect
correspondence between the estimated value and the data, and with 0 indicating no
systematic correspondence whatsoever. In this model the R Square is approximately
97.53%. This is a very high R-square, implying a very good fit.
The analysis can also provide some very revealing graphs: the fit of the regression
to the actual data and the residuals (the difference between the actual and the pre-
dicted values). To produce a residuals plot, check the residuals box in the dialogue
box shown in Exhibit 3.22. This allows you to see the accuracy of the regression
model. In Exhibit 3.23 you can see the Residuals Output at the bottom of the out-
put. The residual for the first observation (23) is 2.857
...
since the predicted value
produced by the regression is 20.143
).
Finally, the coefficients of the regression are also specified in Exhibit 3.23. The Y
intercept or
...
(23–20.143
... =
2.857
...
, 12.2 for the sales data, is where the regression line crosses the Y axis.
The coefficient of the independent variable
α
β
, approximately 7.94, is the slope of the
Search JabSto ::




Custom Search