First, 'cvs update' does not grab the files in lib/gsl-extras. I don't know if I'm doing something stupid, or if some information is missing in CVS. (I can see the files on savannah with a browser.)
I checked gsl's computations for the values shown below (for the beta distribution). I assumed the values in your message were exact (not rounded). cdf.beta(1.00,1.5,.25) should be exactly 1, and this is what gsl_cdf_beta_P(1.00,1.5,.25) gives. pdf.beta(1.00,1.5,.25) should be infinity, and gsl_randist_beta(1.00,1.5,.25) is inf. idf.beta(.99,1.5,.25) is within double precision of 1.00, and this is what gsl_cdf_beta_Pinv(1.00,1.5,.25) gives. It seems gsl is evaluating the relevant functions correctly. I suspect the problem is related to rounding and display of floating point values in PSPP. I'll familiarize myself with PSPP's tests and do some more checking. I did not check the values related to the F distribution yet. -Jason > 225c225 > < .99 1.50 .25 1.00 .99 431322.3 > --- > > .99 1.50 .25 1.00 .98 67836.74 > > (The columns are: > P a b i c p > where i = idf.beta(P, a, b), > c = cdf.beta(i, a, b), > and p = pdf.beta(i, a, b).) > > Here, only the extreme values are off as far as I can tell. I > suppose this must be the root of the problem though. > > (This is a good part of what I was going to bring up with Jason > anyhow.) > -- > "Long noun chains don't automatically imply security." > --Bruce Schneier > > > _______________________________________________ > pspp-dev mailing list > [email protected] > http://lists.gnu.org/mailman/listinfo/pspp-dev > _______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
