[pymvpa] SMLR much slower than LinearCSVMC !?

Yaroslav Halchenko debian at onerussian.com
Thu Apr 26 00:55:07 UTC 2012

SMLR might be faster or slower than SVM depending on data

also lm argument controls sparseness and convergence_tol convergence --
both will have influence on runtime (as few others as well).

but the message: it is ok for it to be slower.  And depending how
installed it you might have got only native Python SMLR implementation
which would be even slower (you should have seen a warning if that was
the case)

On Wed, 25 Apr 2012, Ping-Hui Chiu wrote:

>    Dear PyMVPA experts,

>    Yesterday I changed my classifier from LinearCSVMC to SMLR and the
>    processing speed seems 2-5 times slower. Is this normal? Is there any
>    parameter in SMLR I can set to speed up its process?

>    Thanks!
>    Ping-Hui

> _______________________________________________
> Pkg-ExpPsy-PyMVPA mailing list
> Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa

Keep in touch                                     www.onerussian.com
Yaroslav Halchenko                 www.ohloh.net/accounts/yarikoptic

More information about the Pkg-ExpPsy-PyMVPA mailing list