Mathematica 9 is now available
Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1993
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1993

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

Search the Archive

Normal Distributions in Mma 2.1

  • To: mathgroup at yoda.physics.unc.edu
  • Subject: Normal Distributions in Mma 2.1
  • From: mek at guinan.psu.edu (Mark E. Kotanchek)
  • Date: Mon, 30 Aug 93 22:08:53 -0400

Hi,

I was making sure some arguments I was about to make were, in fact, valid  
involving FFTs of Gaussian signals when I discovered that Mathematica 2.1 on my  
NeXTstation isn't quite accurate. Loading in

   In[2]:=
   Needs["Statistics`ContinuousDistributions`"]
   Needs["ftpAlgebra`Declare`"]
   {Declare, NewDeclare, NonPositive, RealQ}
   


and defining the p.d.f of a normal distribution and checking the distribution  
and the mean leads to

   fx = PDF[NormalDistribution[mu,sigma],x]
   Integrate[fx,{x,-Infinity,Infinity}]
   Integrate[x fx, {x,-Infinity,Infinity}]
   Out[6]=
             2            -2
      Sqrt[mu ] Sqrt[sigma  ] sigma
      -----------------------------
                   mu
   

   Out[7]=
             2            -2
      Sqrt[mu ] Sqrt[sigma  ] sigma
   

which isn't exactly aesthetic, even if it may be technically correct. However,  
if we define x, mu (the mean) and sigma (the standard deviation) to be real and  
restrict sigma to be non-negative, we get,

   Declare[{x, mu, sigma}, Real]
   Declare[sigma,NonNegative]
   Integrate[fx,{x,-Infinity,Infinity}]
   Integrate[x fx, {x,-Infinity,Infinity}]
   Out[10]=
      Abs[mu]
      -------      <----- should be "1"
        mu
   

   Out[11]=
      Abs[mu]      <------ should be "mu"
   

which is clearly wrong since the integral of the p.d.f. is, by definition,  
equal to one and this result would lead to -1 if mu were negative. Similarly,  
the mean should be equal to mu.

OK, I thinks, maybe Mma doesn't like the potential for a standard deviation  
equal to zero. I refine the definitions and, alas, get the same results....

   In[14]:=
   Declare[sigma,Positive]
   Integrate[fx,{x,-Infinity,Infinity}]
   Integrate[x fx, {x,-Infinity,Infinity}]
   Out[13]=
      Abs[mu]
      -------
        mu
   

   Out[14]=
      Abs[mu]
   

I assume that this sort of misbehavior has been reported to WRI. The question  
is whether it has been fixed in Mma 2.2 which should be wandering my way  
eventually?

Mark.
---
Mark Kotanchek
Guidance & Control Dept - 363 ASB
Applied Research Lab/Penn State
P.O. Box 30
State College, PA 16804

e-mail:	mek at guinan.psu.edu (NeXTmail)
TEL:	(814)863-0682
FAX:	(814)863-7843





  • Prev by Date: solution of an equation
  • Next by Date: plotting graphs with asymptotes
  • Previous by thread: solution of an equation
  • Next by thread: plotting graphs with asymptotes