For random variables X and Y, we wish to use Y to estimate X. However, our estimate must be of the form 
= aY
(a) Find a∗, the value of a that minimizes the mean square error e = E[(X - aY)2]
(b) For a = a∗, what is the minimum mean square error e∗?
(c) Under what conditions is the LMSE estimate of X?