On Wed, Jul 18, 2007 at 03:06:27PM +0000, Allin Cottrell wrote:
Update of /cvsroot/gretl/gretl/doc/commands
In directory sc8-pr-cvs2.sourceforge.net:/tmp/cvs-serv29173/doc/commands
Modified Files:
gretl_commands.xml
Log Message:
Gamma p-values: use shape/scale parameterization
Index: gretl_commands.xml
===================================================================
RCS file: /cvsroot/gretl/gretl/doc/commands/gretl_commands.xml,v
retrieving revision 1.264
retrieving revision 1.265
diff -u -d -r1.264 -r1.265
--- gretl_commands.xml 17 Jul 2007 17:29:48 -0000 1.264
+++ gretl_commands.xml 18 Jul 2007 15:06:25 -0000 1.265
@@ -6208,14 +6208,14 @@
given, before the <repl>xval</repl>: for the
<mathvar>t</mathvar>
and chi-square distributions, the degrees of freedom; for
<mathvar>F</mathvar>, the numerator and denominator degrees of
- freedom; for gamma, the mean and variance; for the binomial
- distribution, the <quote>success</quote> probability and the
- number of trials; and for the Poisson distribution, the parameter
- &lgr; (which is both the mean and the variance). Note: the
- parameters for the gamma distribution are sometimes given as shape
- and scale rather than mean and variance. The mean is the product
- of the shape and the scale; the variance is the product of the
- shape and the square of the scale.
+ freedom; for gamma, the shape and scale parameters; for the
+ binomial distribution, the <quote>success</quote> probability and
+ the number of trials; and for the Poisson distribution, the
+ parameter &lgr; (which is both the mean and the variance). Note:
+ the parameters for the gamma distribution are sometimes given as
+ mean and variance rather than shape and scale. The mean is the
+ product of the shape and the scale; the variance is the product of
+ the shape and the square of the scale.
Two notes about this:
- If the user is supposed to feed gretl with the shape and scale parameters, I
think it would be more useful to "take the inverse" of the last sentence,
something like:
The scale parameter is the variance divided by the mean; the shape parameter is
the square of the mean divided by the variance.
(please check the implied maths :-)
- In some textbooks (such as my edition of Mood et al.) the gamma distribution
is parametrized using an inverse-scale parameter (which affects the above
calculations), so we should probably mention which version of the function we
are using. For example, see
http://en.wikipedia.org/wiki/Gamma_distribution
Cri
--
GPG/PGP Key-Id 0x943A5F0E -
http://www.linux.it/~cri/cri.asc
Free software, free society -
http://www.fsfeurope.org