Mathematics desk
< October 22	<< Sep \| October \| Nov >>	October 24 >

Welcome to the Wikipedia Mathematics Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

October 23[edit]

Continuous probability distribution, with finite support, and closed-form quantile function[edit]

For some computational experiments, I need a continuous probability distribution, with finite support, and closed-form quantile function. Kindly list any.

I'm already aware of probability distributions that satisfy my requirements except for finite support: Logistic distribution and Gumbel distribution. Thanks! --RM — Preceding unsigned comment added by 121.247.216.109 (talk) 10:45, 23 October 2014 (UTC)[reply]

How about the triangular distribution? Sławomir Biały (talk) 10:55, 23 October 2014 (UTC)[reply]

Thanks User:Sławomir Biały! I had missed this simple option. Other suggestions are welcome! --BR — Preceding unsigned comment added by 121.247.216.109 (talk) 11:40, 23 October 2014 (UTC)[reply]

It doesn't get simpler than the uniform distribution on [0,1]: Q(p) = p. --Mark viking (talk) 11:44, 23 October 2014 (UTC)[reply]

There's loads of probability distributions at List of probability distributions but yes having a closed-form quantile restricts the choices. I don't normally consider that as a problem with a computer though as given a function getting a good enough inverse should be easy enough even if only by tabulating values in a table and then doing a binary chop and interpolation for actual values.

I presume you mean on a finite interval rather than finite support which sounds more like on a finite set of points. Dmcq (talk) 13:40, 23 October 2014 (UTC)[reply]

Glancing through List of probability distributions#Supported on a bounded interval, the Irwin–Hall distribution, Kumaraswamy distribution, and raised cosine distribution look promising. -- BenRG (talk) 23:32, 23 October 2014 (UTC)[reply]