4.20. Consider a regression through the origin, in =x; +61. With E(Er)=0. lt"l[e,-}=cr2x,v1 x,- :> i}, i: 1.2. . . . .10 a. Derive the generalized (weighted) least squares eslmator of ,6 and obtain its variance. b. Assume that the experimenter recorded only the sample means 3 =15 and? = 30. If possible, obtain a numerical value for the weighted least squares estimate in {a} and express its variance as a function of 0'2