Re: Pareto distribution--need help

From: Roger Bagula (tftn_at_earthlink.net)
Date: 01/26/05

  • Next message: UK Nonlinear News: "UK Nonlinear News: call for material"
    Date: Wed, 26 Jan 2005 16:17:46 GMT
    
    

    npus001@ec.auckland.ac.nz wrote:
    > Hi all,
    >
    > I was doing a simulation based on Pareto distribution. For a set
    > empirical values how can set the values for mode and shape parameter so
    > that the pareto distribution obtained is similar to those empirical
    > values.
    >
    > Thanks
    >
    > Nikhi
    >
    http://mathworld.wolfram.com/ParetoDistribution.html
    http://www.ds.unifi.it/VL/VL_EN/special/special12.html
    f(x)=1-(b/x)^a
    To get the "a" you do a log plot and find the negative slope.
    To get "b" you use the intercept: divide by a and take exp() the result...
    Log[(b/x)^a]=-a*log[x]+a*log[b]
    That's what I get from the formula anyway.
    Roger L. Bagula
    tftn@earthlink.net, 11759Waterhill Road, Lakeside,Ca 92040-2905,tel:
    619-5610814 :
    alternative email: rlbtftn@netscape.net
    URL : http://home.earthlink.net/~tftn


  • Next message: UK Nonlinear News: "UK Nonlinear News: call for material"