[SciPy-user] Stepwise Discriminant Analysis

josef.pktd@gmai... josef.pktd@gmai...
Tue Mar 3 17:49:57 CST 2009


On Tue, Mar 3, 2009 at 4:43 PM, Nuttall, Brandon C <bnuttall@uky.edu> wrote:
> Hello,
>
> Is a stepwise discriminant analysis routine available in Python? Can someone point me in the right direction?
>

I don't know of any stepwise implementation and searching with google
I find mostly articles with the message:

"Stepwise Descriptive or Predictive Discriminant Analysis: Don't Even
Think about Using It!"

is the title of:
http://www.eric.ed.gov/ERICWebPortal/custom/portlets/recordDetails/detailmini.jsp?_nfpb=true&_&ERICExtSearch_SearchValue_0=ED438321&ERICExtSearch_SearchType_0=no&accno=ED438321

or see http://epm.sagepub.com/cgi/content/refs/55/4/525

There are several python packages on machine learning and classifier
analysis that might do something similar to what you want.
e.g. http://www.pymvpa.org/classifiers.html#basic-supervised-learning-methods
or logistic regression in nipy neuroimaging stats.models
FDANode in http://mdp-toolkit.sourceforge.net/tutorial.html#node-list
and there are several others.

However, I don't know how easy it is, to use just the discriminant
analysis without the overhead of learning an entire package, and for
just a linear discriminant analysis they might be overkill.

Josef


More information about the SciPy-user mailing list