6.4.

(a) Relate ML and least squares estimates for the model of Example 6.10.

(b) Show that if the data are iid with the Laplace (double exponential) distribution, f ðyi jyÞ ¼ 0:5 expðj yi  mðxi ÞjÞ; where mðxiÞ ¼ b0 þ b1xi1 þ þ bpxip, then ML estimates of b0; ... ; bp are obtained by minimizing the sum of absolute deviations of the y values from their expected values.