I need to fit a beta-distribution a real data, with a mean of 0 and a standard deviation of 0.17.

I have read that this is possible by using the non-central beta distribution, and I would wlike to scale the standard beta distribution up with 0.33, so that the range is from 0 to 1.33. This new variable should then have a non-central beta distribution.

In R this would look like this:

rbeta(n, shape1, shape2, ncp = 0.33), but shape1(alpha) and shape2(beta) have to be generated from the mean and sd.

Has anyone worked this out in R?

Any help would be much appreciated.

  • 1
    $\begingroup$ It is very hard to determine what you are asking. If you truly "scale up" a Beta variable, its range will no longer be $[0,1]$ and so it cannot have a Beta distribution any more. No Beta distribution has a mean of zero and a positive standard deviation. Infinitely many Beta distributions have a given standard deviation. Since the information you supply is inconsistent, please edit your post either to change the numbers to correct ones or explain what you mean by a "beta distribution," in case it is unconventional. $\endgroup$ – whuber Aug 6 '15 at 12:19
  • $\begingroup$ Thank you for comment. I've edited the original question. Hopefuly this is clearer. $\endgroup$ – user84264 Aug 6 '15 at 15:03

You're overthinking this.

The original range of the data is $[0, 1]$. If you multiply everything by $1.33$, then the data will range from $[0, 1.33]$.


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.