R square and adjusted r square are only applicable

R-square and adjusted-R-square are only applicable for linear regressions, but
not appropriate for nonlinear regressions.
significant depends on the normal distribution assumption.
If
variable selections than the adjusted-R-square does.
Part B – Fill-in Blanks (7 points each)
11. In the two-parameter SLRs, given n = 20, x- bar = 5, β
1
_hat = −1.2, s.e.( β
0
_hat) = 0.35,
= 30 and R
2
= 0.86, calculate σ
2
_hat.
S
yy
12. Consider a general nonlinear regression model f ( X ; γ ), where the matrix X includes 4
explanatory x-variables for n = 50 observations, and the regression parameters γ includes
one intercept-like and 4 slope-like coefficients. Suppose that one uses the Gauss-Newton
Method to estimate regression parameters. Write down the expression of the D-matrix
taught in lectures for the 3
rd
iteration of the method, i.e., one will get g (3) as an update
parameter estimates in the end of this iteration.
of

3