On Fri, 17 Jul 2009 zzh5460(a)yahoo.com wrote:
I got inconsistent results when running PCA using covariance
matrix in gretl. Please see attached TSY_curve file for the data
used. Please see attached PCA_problem file that illustrated the
problem when verified with R.
You are using an old version of gretl. We had a problem with the
ordering of the columns going wrong when printing more than 7
components. This was fixed in gretl version 1.7.8, in September
2008.
See
http://gretl.sourceforge.net/ChangeLog.html for notes on
this and other fixes.
Allin Cottrell