Suppose that we are given i.i.d. random variables X1, ..., X, from some distribution Po where A is the unknown parameter to be estimated. Let 0 = 0(X1, ..., X,) be an estimator of 0 computed from the data X1, ..., Xn. Let Ep denote the expectation to emphasize that the data is from Po. We define three quantities for the estimator 0: The mean squared error (MSE) of 8 is defined as MSE(9, 0) := E,[(0 -9)2]. . The bias of A is defined as Biaso (0) := Eo[0] - 9. The variance of the estimator O is defined as Varo(0) := E[(0 - Eo)2]. Show that the formula of bias-variance trade-off holds: MSE(0, 0) = [Biaso (0)]2 + Varo(@). In particular, if A is unbiased, that is, the bias is zero, then the MSE is equal to the variance