Re: Inverted Bell Curve
- From: Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx>
- Date: Wed, 31 May 2006 11:45:44 -0400
On 30 May 2006 14:16:29 -0700, "Tasci" <google@xxxxxxxxxxxxxxx> wrote:
I'm clear on the fact that the sum of a random sampling from a range of
numbers say, 0 to 4, will distribute itself in the form of the bell
curve. That is, taking more numbers per sample increases the
likelihood of their sum being near the middle. Is there a different
function or relation besides 'sum' that I could use to set up a
situation where the extremes of 0 and 4 have the highest probability,
and the center has the lowest chance of being hit? I want to make it
in this example so that the probability of getting 0 or 4 is very high,
and the probability of getting 2 is very low.
You can tend toward one extreme or the other by taking
the minimum or maximum. Either can be approached as a
generalized mean, using a large positive or negative power.
(Take the nth root of the average of the quantities, x^n)
Instead of the sum, you could define a function, "The furthest
away from 2".
--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.
- References:
- Inverted Bell Curve
- From: Tasci
- Inverted Bell Curve
- Prev by Date: Re: Is there an analog of the central limit theorem for integrals?
- Next by Date: Re: Chi-Squared Test (goodness of fit) confidence levels
- Previous by thread: Re: Inverted Bell Curve
- Next by thread: A question in linear algebra
- Index(es):
Relevant Pages
|
|
