Refer to the form of iterations for the Gauss–Newton algorithm described in Section 11.3.7. Show that an analogous formula

????̂ − ???? = (XTX)

−1XT????

holds for the ordinary linear model, where ???? = y − ???? is a “true residual.”