g

,

, two data sets had different slops. Therefore, the variances of

nt variable were different. They were 29.595 in Figure 4.8(a) and

in Figure 4.8(b). Because of this difference, two linear

n models had different R^ଶ and F-statistic values. For instance, ܴ^ଶ

1 in Figure 4.8(a) and was 0.7739 in Figure 4.8(b). The F-statistic

in Figure 4.8(a) and 1071.5 in Figure 4.8(b). Although two

n models had the same total regression error, their dependent

had different variances. Therefore, a regression model with a

ariance of the dependent variable had lower fitness measurement

d with a regression model with a greater variance of the dependent

although two regression models had the same total regression

(a) (b)

wo regression models for two data sets with a similar total regression error, but

ariances on dependent variable show very different R^ଶ and F-statistic values.

e significance analysis of regression coefficients

question is the significance analysis of the regression coefficients

nship with the importance of the independent variables in a

n model. It examines whether an independent variable is

ntly correlated with the dependent variable. If an independent

has a significant correlation with the dependent variable, the

ent variable is believed to have a significant contribution to the

nt variable. To examine whether an independent variable has a