Recall, that a binomial distribution in which the probability of success at each trial is randomly drawn from a beta distribution results in the so called beta-binomial distribution. One can calculate the probability mass function explicitly as

$$p_{BB(n, \alpha, \beta)}(k) := \mathbb{P}[X=k \mid n, \alpha, \beta] = \binom{n}{k} \frac{B(\alpha + k, \beta + n-k)}{B(\alpha, \beta)} \; .$$

Now, following up this answer I find myself in the following situation:

Given $n$ Bernoulli trials, such that the success rate for each trial is modeled with a distribution $p \sim(U-L)X_B + L$, where $X_B\sim Beta(\alpha, \beta)$ is beta distributed and $0 \leq L < U \leq 1$ are constants. What is the corresponding probability mass function for this modified beta-binomial distribution?

probabilityislogic mentioned in his post that

"[...] noting that making the change of variables $p=\frac{x-L}{U-L}$ for $L<x<U$ the beta integral transforms to:" $$B(\alpha,\beta)=\int_{L}^{U}\frac{(x-L)^{\alpha-1}(U-x)^{\beta-1}}{(U-L)^{\alpha+\beta-1}}dx$$

But I'm not entirely sure how to use that information to get the desired pmf. More specifically, I wasn't able to deduce an expression of $p_{BB(L, U, n, \alpha, \beta)}$ in terms of $p_{BB(n, \alpha, \beta)}$.

Cross Post:
This is a cross post from Cross Validated SE. I am aware that doing this is suboptimal but decided that it might be appropriate in this case.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.