

The program data.for will read the file temp.ini and
write out two files that can be used as input to the
difference in means program.  The temp.ini file has
the following content:
  50   55 4 -1
  50   50
  5.48   5.48


C
       READ(8,*)N1,N2,ICOUNT,ISEED
       READ(8,*)C1,C2
       READ(8,*)S1,S2
C

Only the first three lines are read the remaining lines
are there as reminders to me as to what these inputs are.
The first line is the number of points in data set 1,
number in set 2, icount is just an index that says, in
this case, skip four random numbers and then use the fifth.
Its just a way of making the random number generator a
little more random.  The last number, iseed, here -1,
is a seed for the gaussian random number generator.
It can be any negative number and different seeds
generate different random sequences.

The second line are the two means to generate,
and the third are the two standard deviations.

The data in this directory may be used to test the programs
to make sure everything is working correctly.  If you look
at the files suffixed with the word output, you will find:

 The First Input File Is:  bayes.1
 The Second Input File Is: bayes.2

 Enter The Amplitude Lower Bound: 30.00
 Enter The Amplitude Upper Bound: 100.0

 Enter The Variance Lower Bound: 3.000
 Enter The Variance Upper Bound: 8.000

 Enter the number of steps to take in going
   from the lowest difference to the highest (max 1001): 51

where these lines contain the information that was entered
to make a given run.  You should be able to enter these
values and get the identical results; if  you can't the
programs did not compile correctly; or something else
is in error -- regardless if you cannot reproduce the runs
let me know.

I am not going to go over the rest of the output files,
they are explained in the article.  The remaining files,
the have the suffix of 1, 2, delta, and ratio.
The suffix 1 and 2 are the data files.  These are ascii, single
column, and they can be integers, reals, or floats -- one
per line.  The ratio and delta files are the probability
distributions for the ratio of the variances and the
difference in means.  Here is a sample of the delta files:

       del       P(del|same DI) P(del|diff DI)   P(del|DI)
   -6.32389D-01    1.47244D-01    1.47255D-01    1.47248D-01
   -3.52389D-01    1.98646D-01    1.98650D-01    1.98648D-01
   -7.23890D-02    2.52111D-01    2.52106D-01    2.52109D-01
    2.07611D-01    3.00747D-01    3.00733D-01    3.00741D-01
    4.87611D-01    3.37003D-01    3.36982D-01    3.36994D-01
    7.67611D-01    3.54575D-01    3.54551D-01    3.54565D-01
    1.04761D+00    3.50224D-01    3.50200D-01    3.50214D-01
    1.32761D+00    3.24762D-01    3.24743D-01    3.24754D-01
    1.60761D+00    2.82804D-01    2.82793D-01    2.82800D-01
    1.88761D+00    2.31380D-01    2.31378D-01    2.31379D-01
    2.16761D+00    1.77990D-01    1.77997D-01    1.77993D-01

where "del" is the difference in means, "same" means the same
variance, "DI" is the data and the prior information, "diff"
means different variances -- so P(del|same DI) is the probability
for the difference being del given the same variances, the
data and the prior information; etc., for the others.

The ratio files look as follows:

      rat         P(rat|same DI) P(rat|diff DI)   P(rat|DI)
    5.86478D-01    1.33417D-01    1.35654D-01    1.33532D-01
    6.09083D-01    2.75196D-01    2.79311D-01    2.75406D-01
    6.31688D-01    4.98558D-01    5.04888D-01    4.98881D-01
    6.54293D-01    8.11769D-01    8.20045D-01    8.12192D-01
    6.76898D-01    1.20841D+00    1.21764D+00    1.20888D+00
    6.99503D-01    1.66555D+00    1.67421D+00    1.66599D+00
    7.22108D-01    2.14586D+00    2.15227D+00    2.14619D+00
    7.44713D-01    2.60347D+00    2.60625D+00    2.60362D+00
    7.67318D-01    2.99223D+00    2.99065D+00    2.99215D+00
where here P(rat|same DI) means the probability for the ratio
of the variances given the "same" means and the data, "D", and
the prior information.

Needless to say these two files were designed for making plots.
Any plots you want to make will have to be done on your own.

If you have questions or comments please contact me.
 
Larry Bretthorst
Dept. of Chemistry
Washington University
St. Louis Mo., 64130

Email: larry@bayes.wustl.edu
       larry@wuchem.wustl.edu
Phone: 314-362-9994 (work)
