PRML exercise 1.8:
To derive mean: change of variable z = x - u, use symmetry
To derive variance: differentiate normaliation condition (integrate to 1) w.r.t. sigma^2.
