+ Reply to Thread
Results 1 to 3 of 3

Thread: mean and variance of a beta distribution

  1. #1
    Points: 632, Level: 12
    Level completed: 64%, Points required for next Level: 18

    Posts
    24
    Thanks
    8
    Thanked 2 Times in 2 Posts

    mean and variance of a beta distribution




    I'm trying to do point (b) of exercise 3.30 of the book "Statistical Inference" (Casella & Berger).
    The exercise says:
    "Use the identities of Theorem 3.4.2. to
    (a) calculate the variance of a binomial random variable.
    (b) calculate the mean and variance of a beta(a,b) random variable."
    Theorem 3.4.2 says the following.
    If X is a random variable with pdf or pmf of the form
    f(x|\theta) = h(x)c(\theta)\exp\left( \sum_{i=1}^k w_i(\theta)t_i(x) \right)
    (exponential family), then
    E \left( \sum_{i=1}^k \frac{\partial w_i(\theta)}{\partial \theta_j} t_i(X) \right) = - \frac{\partial}{\partial\theta_j}\log c(\theta)
    and
    Var \left( \sum_{i=1}^k \frac{\partial w_i(\theta)}{\partial \theta_j} t_i(X) \right) = - \frac{\partial^2}{\partial\theta_j^2}\log c(\theta) - E \left( \sum_{i=1}^k \frac{\partial^2 w_i(\theta)}{\partial \theta_j^2} t_i(X) \right).

    Point (a) is doable, but what about point (b)? First of all, since we have two parameters (a and b), do we get a system of two equations? My difficulties is that I have a cumbersome E[logX] and then I have to take derivatives of B(a,b). It's a nightmare... If I'm not wrong,
    c(a,b) = 1/B(a,b)
    t_1(x) = \log x, w_1(a,b) = a-1
    t_2(x) = \log (1-x), w_2(a,b) = b-1.
    Any help?

  2. #2
    TS Contributor
    Points: 18,866, Level: 87
    Level completed: 4%, Points required for next Level: 484

    Posts
    2,831
    Thanks
    10
    Thanked 517 Times in 492 Posts

    Re: mean and variance of a beta distribution

    In my Casella and Berger, part b is asking for \text{Poisson}(\lambda) instead of Beta distribution. Not sure if this method is useful for Beta distribution. Anyway you may take a look at

    http://en.wikipedia.org/wiki/Beta_di...Geometric_mean

    and there are lots of calculations for E[\ln X] and other related terms. And as you expected you will need to in terms of some non-elementary function like the digamma function which is the derivative of log-gamma function.

  3. The Following User Says Thank You to BGM For This Useful Post:

    Kiuhnm (10-12-2013)

  4. #3
    Points: 632, Level: 12
    Level completed: 64%, Points required for next Level: 18

    Posts
    24
    Thanks
    8
    Thanked 2 Times in 2 Posts

    Re: mean and variance of a beta distribution


    Thank you for your answer. I've had a look at the errata and this is indeed an error: part b should say Poisson!

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats