i.pca

GRASS GRASS user
1

igny-Giroux??at?Nov?22??95?04?03?26?pm@bnr.ca>
content-type: text/plain; charset=US-ASCII
organization: University of Illinois at Urbana
mime-version: 1.0
reply-to: grassu-list@max.cecer.army.mil
newsgroups: info.grass.user
originator: daemon@ux1.cso.uiuc.edu

Does anyone know whether i.pca uses a correlation OR covariance matrix to
generate the corresponding components? There is no reference to this in the
manual and the results from r.covar and m.eigensystem are dramatically
different depending on whether a correlation or covariance matrix was used.

Thanks.
L.A. Dupigny-Giroux

Please send replies to this list, as I'm sure others are interested too.

        Martijn

P. Martijn van Leusen -------------------------------------------------------+
Leverhulme Research Fellow, The Wroxeter Hinterland Project |
University of Birmingham Field Archaeology Unit |
Edgbaston, Birmingham B15 2TT Phone +44 121 414 5513 Fax +44 121 414 5516 |
E-mail P.M.van-Leusen@bham.ac.uk |
URL http://www.bham.ac.uk/BUFAU/Staff/vanleupm.html__________________________|

2

"" ..and the results from r.covar and m.eigensystem are dramatically

different depending on whether a correlation or covariance matrix was used.""

As it has to be. If I recall correctly, i.pca uses the covar matrix. If
you are using m.eigensystem it's not difficult to check out: use r.mapcalc
and the eigenvectors.

Agus

***************************************************
Dr. Agustin Lobo
Instituto de Ciencias de la Tierra (CSIC)
Marti Franques s/n
08028 Barcelona (Spain)

tel. 34 3 330 2716
fax. 34 3 411 0012

alobo@ija.csic.es
alobo@eno.princeton.edu
***************************************************