Consider the model: y=β0+β1X1+β2X1²+β3X2+β4X1X2+u, where E[u|X1,X2]=0 and Var(u)= σ²(X1,X2)
prepare down a WLS-transformed version of the model that has a homoskedastic error term. Then verify that your transformed model satisfies the zero conditional mean assumption and that it is homoskedastic; i.e. demonstrate that the transformed equation satisfies MLR.4 and MLR.5..