Re: PSPP-BUG: pspp bug

[John Darrington]

OK, I had a look at this closer.   The root cause is that the
example data are very nearly linearly correlated.  Hence the
correlation matrix has a determinant close to zero and thus, the
KMO and Bartlett measures are very unstable.

Since they don't have any relevance to the bug being tested, for 
now I have simply removed them from the test.


J'

On Thu, Apr 05, 2012 at 10:12:53PM -0700, Ben Pfaff wrote:
     When we looked at this in IRC, I think we were looking at the
     wrong value.  You were talking about log(0) but that's involved
     in "Bartlett's Test of Sphericity" not in "Kaiser-Meyer-Olkin
     Measure of Sampling Adequacy".
     
     I added
           printf ("sum_ssq_r=%e sum_ssq_a=%e\n", sum_ssq_r, sum_ssq_a);
     as line 1870 of factor.c and got:
             sum_ssq_r=3.472064e+35 sum_ssq_a=1.036522e-30
     so that the calculation sum_ssq_r /  (sum_ssq_r + sum_ssq_a) is
     basically x / (x + 0) == 1, which seems reasonable.  What do you
     see when you add that line?
     
     John Darrington <address@hidden> writes:
     
     > I'll take a look at it.
     >
     > On Wed, Apr 04, 2012 at 09:17:17AM -0700, Ben Pfaff wrote:
     >      That seems bad.  "1.00" and NaN are pretty different.  Do you
     >      think it's worth looking into the calculations, or should we just
     >      ignore this difference as we have others?
     >      
     >      John Darrington <address@hidden> writes:
     >      
     >      > My system fails the new test:
     >      >
     >      > 395. factor.at:1803: testing FACTOR /ROTATION=NOROTATE ...
     >      > ../../tests/factor.at:1821: pspp -O format=csv factor-norotate.sps
     >      > --- -   2012-04-04 11:18:24.388115600 +0100
     >      > +++ 
/home/Darringt/.p/P/UTP_00/ClearQuest/tests/testsuite.dir/at-groups/395/stdout  
    2012-04-04 11:18:23.366814100 +0100
     >      > @@ -7,7 +7,7 @@
     >      >  TRAIT5,7.29,2.66,7
     >      >  
     >      >  Table: KMO and Bartlett's Test
     >      > -Kaiser-Meyer-Olkin Measure of Sampling Adequacy,,1.00
     >      > +Kaiser-Meyer-Olkin Measure of Sampling Adequacy,,NaN
     >      >  Bartlett's Test of Sphericity,Approx. Chi-Square,NaN
     >      >  ,df,10
     >      >  ,Sig.,NaN
     >      > 395. factor.at:1803:  FAILED (factor.at:1821)
     >      >
     >      >
     >      > On Tue, Apr 03, 2012 at 10:00:28PM -0700, Ben Pfaff wrote:
     >      >      Thank you very much for the bug report.  I reproduced the 
bug and
     >      >      fixed this on the PSPP "master" branch.
     >      >      -- 
     >      >      Ben Pfaff 
     >      >      http://benpfaff.org
     >      >      
     >      >      _______________________________________________
     >      >      Bug-gnu-pspp mailing list
     >      >      address@hidden
     >      >      https://lists.gnu.org/mailman/listinfo/bug-gnu-pspp
     >      
     >      -- 
     >      Ben Pfaff 
     >      http://benpfaff.org
     
     -- 
     Ben Pfaff 
     http://benpfaff.org

-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://keys.gnupg.net or any PGP keyserver for public key.

signature.asc
Description: Digital signature