Professor May B. Hard, who has a tendency to give difficult problems in probability quizzes, is concerned about one of the problems she has prepared for an upcoming quiz. She therefore asks five TAs to solve the problem and record the solution time.
May's prior probability that the problem is difficult is 0.3, and she knows from experience experience that the conditional PDF of her TA's solution time X, in minutes, is
fX|Θ(x|Θ = 1) =
c1e-0.04x, if 5 ≤ x ≤ 60,
0, otherwise
if Θ = 1 (problem is difficult) and
fX|Θ(x|Θ = 2) =
c2e-0.16x, if 5 ≤ x ≤ 60,
0, otherwise
if Θ = 2 (problem is not difficult).
where c1 and c2 are normalizing constants.She uses the MAP rule to decide whether the problem is difficult
1. Given that the TA's solution time was 20 minutes, which hypothesis will she accept and what will be the probability of error?
2. Not satisfied with the reliability of her decision, May asks four more TAs to solve the problem. The TAs' solution times are conditionally independent and identically distributed with the solution time of the first TA. The recorded solution times are 20,
10, 25, 15, and 35 minutes. On the basis of the five observations, which hypothesis will she now accept, and what will be the probability of error?