Consider the nonlinear regression model y = x 1 + / ( x 2 ) + u , where x 1

Consider the nonlinear regression model $y=\alpha x_1+\gamma /(x_2-\delta )+u$, where $x_1$ and $x_2$ are exogenous regressors independent of the iid error $u\sim \mathcal{N}[0,{\sigma }^2]$.

(a) Derive the equation for the Gauss-Newton algorithm for estimating $(\alpha ,\gamma ,\delta )$.

(b) Derive the equation for the Newton-Raphson algorithm for estimating $(\alpha ,\gamma ,\delta )$.

(c) Explain the importance of not arbitrarily choosing the starting values of the algorithm.