Let P be a uniform random variable on the interval $(0,1)$ with density function f(p) = 1, $0<p<1$. Let $X_i|P$, i = 1,2,...,n be independent and identically distributed random variables having a Bernoulli distribution with parameter P, where $P(X_i=1|P=p) = p$ and $P(X_i=0|P=p) = 1-p$. Are $X_1,X_2,...,X_n$ exchangeable? Are they independent?

The hint is the problem is a Gamma distribution.

  • $\begingroup$ What is your $f(x\mid p)$ supposed to represent? $\endgroup$ – Did Sep 23 '14 at 16:08

Take any n tuple $x$ of 0's and 1's with k 1's and n-k 0's. Then $$ P((X_1,\ldots, X_n) = x|P=p) = p^k(1-p)^{n-k} $$ So, $$P((X_1,\ldots, X_n) = x) = \int_0^1p^k(1-p)^{n-k}dp = \frac{\Gamma(k+1)\Gamma(n-k+1)}{\Gamma(n+2)}$$

Since the above expression only depends upon the number of $1$'s, you can permute them and obtain the same probability. This shows exchangeability.

But $P(X_1 = 1) = P(X_2 = 1) = \int_0^1pdp = 1/2$. But $$P(X_1=1,X_2=1) = \frac{\Gamma(3)}{\Gamma(4)} = 1/3$$

so they are not independent!


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.