Question 1 - Suppose we have n independent observations from
P(y|θ) = θ/y2, y > θ, and θ > 0
(a) Find a conjugate prior distribution for θ.
(b) Find the posterior mean and variance for θ.
(c) Suppose n = 3 and observations are 2; 5 and 7. Using simulation, find the posterior mean and variance for θ and compare your results with (b).
(d) Using simulation and the observations in (c), find a 95% credible set for θ.
Question 2 - Suppose that your prior for θ is a ¼ to ¾ for a mixture of N(0, 1) and N(1, 1) and that a single observation y∼N(θ, 1) turns out to equal 3. What is your posterior probability θ > 2?