Jeffreys' prior: For the multivariate normal model, Jeffreys' rule for gen- erating a prior distribution on (0, _) gives py (0, _) x |_|-(P+2)/2. a) Explain why the function py cannot actually be a probability density for (0, E). b) Let py(0, Ely , ... . . ) be the probability density that is proportional to py(0, E)xp(y1; . .., y,|0, _). Obtain the form of py(0, Ely1, . . . . yn), PJ( 0 5, y . . . ., y ) and py (C y . . . ., y.)