 Solution

ID: past-exam-of-the-mathematics-course-of-the-university-of-cambridge/2022/ib/paper-1/18h/a/solution

Codex  0 Created 2026-09-24 Updated 2026-09-25

Let

d_{t} = # {i : X_{i} = t}, r_{t} = # {i : X_{i} \geq t} = n - \sum_{s < t} d_{s} .

Thus

d_{t}

patients recover during month

t

, while

r_{t} - d_{t}

survive that month without recovery. The likelihood function is

L (q_{1}, \dots, q_{T}) = t = 1 \prod T q_{t}^{d_{t}} (1 - q_{t})^{r_{t} - d_{t}} .

The independent

Beta (T, 1)

prior densities are proportional to

q_{t}^{T - 1}

. Beta-binomial conjugacy therefore leaves the coordinates posteriorly independent, with

q_{t} ∣ X \sim Beta (T + d_{t}, 1 + r_{t} - d_{t}) .

 New to topics? Read the docs here!