s-news
[Top] [All Lists]

Re: Estimating distributions

To: "'David M Smith'" <dsmith@insightful.com>, s-news@wubios.wustl.edu
Subject: Re: Estimating distributions
From: "Gunter, Bert" <bert_gunter@merck.com>
Date: Wed, 27 Jun 2001 08:38:39 -0400
The following post by David Smith to this list prompts what seems to be an
obvious question:

*******************************
-----Original Message-----
From: David M Smith [mailto:dsmith@insightful.com]
Sent: Monday, June 25, 2001 1:01 PM
To: s-news@wubios.wustl.edu
Cc: P.J.Wells@open.ac.uk
Subject: Re: [S] Estimating distributions


The EnvironmentalStats module for S-PLUS has capabilities for estimating
parameters from a wide variety of distributions, using methods including MLE
and Method of Moments.  For example, it adds a function "enorm" (e is for
"estimate") to dnorm, rnorm, etc., and similar functions for other
distributions in S-PLUS and more besides.

Despite the name, EnvironmentalStats is useful in just about all areas of
statistics.
*************************************

If this is the case, why isn't this functionality already in S-Plus? Is it
sensible marketing to say: "We have additional features that almost all
users would find useful but which you must pay extra for?"  I would have
thought that "basic" features of interest to almost all users would be in
the base product. Special modules with special features for subsets of users
would be extra-cost add-ons.

(Disclaimer: Personal opinion only that does not in any way reflect the
views of my company or my colleagues).

Bert Gunter
mailto: bert_gunter@merck.com

"The business of the statistician is to catalyze the scientific learning
process."      -- George E.P. Box

<Prev in Thread] Current Thread [Next in Thread>