a. Consider the linear regression model (expressed in matrices notation): Y = XB+E where c has mean 0 and variance-covariance matrix E, which is a known non-singular symmetric positive definite matrix. In book, and also in class, the estimator of B obtained by the general least squares method is B = (XE- X)XC-Y. Show that S is an unbiased estimator of 8 and var(B) = (X'S-1X)-1