 Solution

ID: past-exam-of-the-options-i-course-of-the-university-of-oxford/2013/o/2/a/solution

Let

s = \sum_{j = 1}^{n} x_{j}

. The Bernoulli distribution likelihood is proportional to

θ^{s} (1 - θ)^{n - s}

π (θ ∣ x) \propto θ^{a + s - 1} (1 - θ)^{b + n - s - 1} .

Normalizing with the beta function therefore gives

θ ∣ x \sim Beta (a + s, b + n - s), π (θ ∣ x) = \frac{θ ^{a + s - 1} ( 1 - θ ) ^{b + n - s - 1}}{B ( a + s , b + n - s )} .

The parameters remain positive. This is Beta-binomial conjugacy: successes add to the first shape parameter and failures to the second.

 New to topics? Read the docs here!