Consider a sequence of independent bernoulli random variables $X_1 X_2 ... X_n$ with parameter $\theta$ and $0<\theta<1$,where $P(X_i=1)=1-P(X_i=0)=\theta$. Assume the prior of $\theta$ follows Beta(3,2). Given that the first 9 observations are such that $\sum_{X_i = 1}^9 X_i = 7 $. what would be the (Bayesian) probability that $X_{10} =1$?

So I find out the posterior probability...

$P( \theta |X) \propto \theta^9(1-\theta)^3 $

And question ask me that $P(X_{10} = 1| H=7, T=2)$

But I don't have any idea for solving this question...



  • Your posterior distribution for $\theta$ with density proportional to $\theta^9(1-\theta)^3$ is another Beta distribution. You should try to find its parameters

  • The posterior probability that the next value will be $1$ is then equal to the posterior expected value of $\theta$

  • One approach (not the quickest if you are familiar with Beta distributions) might be to find $$\dfrac{\int\limits_0^1 \theta \cdot \theta^9(1-\theta)^3\, d \theta}{\int\limits_0^1 \,\,\,\,\,\,\, \theta^9(1-\theta)^3\, d \theta}$$

  • $\begingroup$ How about the next value will be 0? Is it still the expected value of theta? $\endgroup$ – Simon s Park Oct 21 '18 at 0:06
  • $\begingroup$ No - the probability that the next value will be $0$ is equal to the expected value of $1-\theta$ $\endgroup$ – Henry Oct 21 '18 at 0:08

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.