********************************************************************
** macros.readme **
********************************************************************
The two Minitab programs 'p_disc' and 'p_disc_p' will compute posterior
probabilities and predictive probabilities, respectively, for a
proportion using a discrete prior distribution. Section 2 describes
the basic formula used by these programs, and Section 3.1 describes an
application of the program 'p_disc_p' in predicting the number of home
runs in future at-bats of a baseball player.
These Minitab 'exec' macros are written for Release 10 of Minitab
(Windows and Macintosh). A second version of the two programs has been
written that will run on older releases of Minitab with character
graphics. Contact Jim Albert at albert@math.bgsu.edu for copies of
these programs.
The program 'p_disc' will compute posterior probabilities for a
proportion. The program assumes that the values of the proportion are
placed in a column 'p' and the associated prior probabilities in the
column 'prior'. Then the program is run by typing
exec 'p_disc'
One inputs the number of successes (s) and number of failures (f) of
the binomial experiment. The output is the posterior probability
distribution for the proportion and a line graph of this distribution.
The posterior probabilities are stored in the column 'post'.
The program 'p_disc_p' will compute predictive probabilities. The
program assumes that the values of the proportion are placed in the
column 'p' and the corresponding probabilities are in either the column
'prior' or the column 'post'. Run the program by typing
exec 'p_disc_p;
One inputs the location of the probabilities (in 'prior' or 'post'),
the sample size of the future experiment, and the range of values (low
and hi) for which predictive probabilities are to be computed. The
output is the predictive distribution for the values of interest and a
line graph of this distribution. The values and probabilities of the
predictive distribution are stored in the columns 'succ' and 'pred'.
The programs 'p_disc' and 'p_disc_p' are included in a package of
Minitab macros for teaching Bayesian inference at an elementary level.
The programs will be included in the book _Bayesian Computation Using
Minitab_, to be published by Wadsworth in the spring of 1996.
_______________________________________________________________________