Mathematica 9 is now available
Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2008

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Normal Probability plot

  • To: mathgroup at smc.vnet.net
  • Subject: [mg91700] Re: Normal Probability plot
  • From: P_ter <petervansummeren at gmail.com>
  • Date: Sat, 6 Sep 2008 02:05:09 -0400 (EDT)

I mostly check the model (in this case the normal distribution) against the data with Quantiles:
 myq = 100000;
data = Sort@RandomReal[NormalDistribution[5, 1], { myq}];
ListPlot[Transpose@{Quantile[NormalDistribution[0, 1], (Range@ myq - .5)/ myq],data}, Frame -> True, Axes -> None]
The ListPlot looks beautiful, but!
First I make a list of the result of FindFit:
varmed = FindFit[Transpose@{Quantile[NormalDistribution[0, 1],(Range@ myq - .5)/ myq], data},a x + b, {a, b}, x]
myy = Quantile[(data - varmed[[2, 2]])/varmed[[1, 2]], (Range@ myq - 0.5)/ myq];
myx = Quantile[NormalDistribution[0, 1], (Range@ myq - .5)/ myq];
mp = Transpose@{myx, myy - myx};
ListPlot[mp]
One can observe how the quantiles of the data (no sample correction) fit the quantiles of the model.
In the above example they both come from the normal distribution.
with friendly greetings,
P_ter


  • Prev by Date: phase-space versus controlling parameter surface
  • Next by Date: An Argument Problem
  • Previous by thread: Re: phase-space versus controlling parameter surface
  • Next by thread: Re: Re: Normal Probability plot