Re: Algorithm for approximation of normal distribution



Gus Gassmann wrote:
On Jan 23, 9:51 am, thomas <tohenn...@xxxxxxxx> wrote:
Hello everyone,

I am looking for an algorithm to approximate the probability function
of a normal distribution. Or strictly speaking, I want to approximate
a single p-value for a given z-value.

As we know, there exists no integral of the normal distribution.

I'd like to address this common misconception. Of course the integral
exists. However, it is not representable by elementary functions, so
the evaluation of the integral can only be done numerically.


.... and there is code around to evaluate such things ... the OP might start here: http://lib.stat.cmu.edu/apstat/
or http://users.bigpond.net.au/amiller/
or look more widely on netlib: http://www.netlib.org/

David Jones

.



Relevant Pages

  • Algorithm for approximation of normal distribution
    ... I am looking for an algorithm to approximate the probability function ... a single p-value for a given z-value. ... there exists no integral of the normal distribution. ... In a book about statistics I found the citation of a paper by Sletten, ...
    (sci.stat.math)
  • Re: Algorithm for approximation of normal distribution
    ... Or strictly speaking, I want to approximate ... a single p-value for a given z-value. ... there exists no integral of the normal distribution. ... In a book about statistics I found the citation of a paper by Sletten, ...
    (sci.stat.math)
  • Re: Algorithm for approximation of normal distribution
    ... Or strictly speaking, I want to approximate ... a single p-value for a given z-value. ... there exists no integral of the normal distribution. ... In a book about statistics I found the citation of a paper by Sletten, ...
    (sci.stat.math)
  • Re: FFT test with few kbits
    ... > two different Fourier components are actually what ... > dependence between two Fourier components becomes ... Strictly speaking, we know that the components are not independent, but this ... for the normal distribution. ...
    (sci.crypt)
  • Re: Limits of log normal distribution
    ... Taking that average of the two range extremes is ... a very poor way of approximating the mean of a normal distribution, ... "Continuous Univariate Statistics, vol. 1" does give a variety of methods ... probability function. ...
    (sci.stat.math)