
VALIDATION THROUGH STATISTICAL ANALYSIS
Correlation Coefficient and Coefficient of Determination
As mentioned earlier, one way of validating if the polynomial equation represents
a good fit is through the coefficient of determination.
First, we determine the sum of squares around the mean for the dependent
variable (y) and label the sum as S
t.
This is the amount of spread in the dependent
variable that exists prior to regression. After performing the regression, we
compute for S
r
, which is the sum of squares of the residual around the regression
curve. This represents the spread that remains after regression.
∑
−=
2
yyS
t
(2.15)
−=
2
er
yy