[Bioc-devel] MLInterfaces changes

Vincent Carey 525-2265 stvjc at channing.harvard.edu
Tue Sep 11 07:58:10 CEST 2007



I have added a README.txt in the top of MLInterfaces, describing the
goals of MLInterfaces 2.0.  We are at 1.11.15 and I hope to get to 2.0
in the next week.

Briefly, the aims are

a) all activities occur through MLearn( formula, data, learnerSchema, numeric, ...) or
MLearn( formula, data, learnerSchema, xvalSpec, ...).  no more separate xval or xvalML or *B
methods

b) MLearn always returns an instance of classifierOutput or clusteringOutput; no more
vectors returned by xval or xvalML

c) the xvalSpec instance can include a general function to define the partition, and
a function that selects features for each partition; feature selection history
is retained

Full back-compatibility is currently maintained, so the *B and xval and old MLearn/xvalML
functionalities are unchanged -- the new MLearn just involves a new signature.  I will
only deprecate when comparable facilities are well-established in the new code base.

example(MLearn) executes a long list of applications including one with formula+ExpressionSet,
balanced cross-validation partition, and feature selection

---
Vince Carey, PhD
Assoc. Prof Med (Biostatistics)
Harvard Medical School
Channing Laboratory - ph 6175252265 fa 6177311541
181 Longwood Ave Boston MA 02115 USA
stvjc at channing.harvard.edu



More information about the Bioc-devel mailing list