[pymvpa] _clf_internals question

Yaroslav Halchenko debian at onerussian.com
Sat Apr 25 17:42:56 UTC 2009


Hi Per,

Thanks!

1.
would you mind sharing the pdf for the paper
http://www-stat.stanford.edu/~hastie/Papers/glmnet.pdf

since I can't download it -- pdf comes broken (tried from 2 computers)

2.
> > Should I make two other instances of the class, similar to what is
> > done for SVMs, that separate this out: i.e., a GLMNET_Reg and
> > GLMNET_Class for regression and classification?
> This seems to be the easiest way.

yeah -- although it possible to do in a fashion similar to the one we
have in SVMs (where it could be SVM or SVR)... it is a bit convolved but
logic is present in mvpa/clfs/_svmbase.py  and a specific kind depends
on the svm_impl parameter to constructor and corresponding
_KNOWN_IMPLEMENTATIONS entry in specific implementation like
mvpa/clfs/sg/svm.py

But since you have implemented it already -- it is cool as is. I would
just either fully abbreviate as 

GLMNETC and GLMNETR (may be with _), so it is in the same fashion as
some other names like GPR

or fully extend as
GLMNET_Classifier and GLMNET_Regression

just to be explicit

-- 
Yaroslav Halchenko
Research Assistant, Psychology Department, Rutgers-Newark
Student  Ph.D. @ CS Dept. NJIT
Office: (973) 353-1412 | FWD: 82823 | Fax: (973) 353-1171
        101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102
WWW:     http://www.linkedin.com/in/yarik        



More information about the Pkg-ExpPsy-PyMVPA mailing list