> > 3, Running .x principal.C, I get:
> > root [0] .x principal.C
> > *************************************************
> > * Principal Component Analysis *
> > * *
> > * Number of variables: 10 *
> > * Number of data points: 10000 *
> > * Number of dependent variables: 3 *
> > * *
> > *************************************************
> > Error in <TPrincipal::MakeEigenVectors>: too many iterations
> >
> > Variable # | Mean Value | Sigma | Eigenvalue
> > -------------+------------+------------+------------
> > 1 | 4.998 | 1.002 | nan
> > 2 | 8.022 | 2.816 | nan
> > 3 | 1.992 | 1.987 | nan
> > 4 | 5.008 | 0.9893 | nan
> > 5 | 8.045 | 2.811 | nan
> > 6 | 2.019 | 2.049 | nan
> > 7 | 4.99 | 1.008 | nan
> > 8 | 1.751e+05 | 1.013e+05 | nan
> > 9 | nan | nan | nan
> > 10 | 1.401e+05 | 8.108e+04 | nan
Hi,
this seems to be caused by an uninitialized array element in
principal.C. Namely data[8] is accessed before its value being set in
"// Then we create the correlated variables" loops.
Jiri
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:46 MET