Third Workshop on
ROC Analysis
in ML
Pittsburgh, USA, 29 June, 2006
held within
ICML’2006,
the 23rd International Conference on
Machine Learning
Brief
description
Receiver Operating Characteristic Analysis (ROC Analysis) is related
in
a direct and natural way to cost/benefit analysis of diagnostic
decision making. Widely used in medicine for many decades, it has been
introduced relatively recently in machine learning. In this context,
ROC analysis provides tools to select possibly optimal models and to
discard suboptimal ones independently from (and prior to specifying)
the cost context or the class distribution. Furthermore, the Area Under
the ROC Curve (AUC) has been shown to be a better evaluation measure
than accuracy in contexts with variable misclassification
costs and/or imbalanced datasets. AUC is also the standard measure when
using classifiers to rank examples, and, hence, is used in applications
where ranking is crucial, such as campaign design, model
combination, collaboration strategies, and co-learning.
Nevertheless, there are many open questions and some limitations that
hamper a broader use and applicability of ROC analysis. Its use in data
mining and machine learning is still below its full potential. An
important limitation of ROC analysis, despite some recent progress, is
its possible but difficult extension for more than two classes.
This workshop follows up a first workshop (ROCAI'04) held
within
ECAI-2004 and a second workshop (ROCML'05) held
within ICML-2005.
This third workshop is intended to investigate on
the hot topics identified during the two previous workshops (e.g.
multiclass
extension, statistical analysis, alternative approaches), on the one
hand, and
to encourage cross-fertilisation with ROC practitioners in medicine, on
the
other hand, thanks to an invited medical expert.
Dr. Darrin C. Edwards, Department of Radiology, University of Chicago
will present the state of the art of ROC analysis in
radiology. Moreover a three-class medical dataset is available to
support exchanges between medical experts and participants. We
strongly encourage submissions dealing with the provided medical
dataset. The three-class
medical dataset will be available on this webpage soon. Please send an
email at lachiche@lsiit.u-strasbg.fr
if you want to join the mailing list where updates will be sent.
Organisation
We encourage submissions on hot topics raised during previous
editions, e.g. ROC analysis software repository, multiclass extension,
statistical analysis. To promote a
workshop atmosphere the program committee will select a few topics and
relevant papers. The program and accepted papers will be published on
the workshop website before the call for participation where we will
encourage people to read them in order to focus on the discussion
during the workshop. The authors of accepted papers will be asked to
prepare the discussion rather than to detail their paper, by
summarising their contributions, on the one hand, and by comparing them
to the other points of view presented in their session, on the other
hand.
The attendance to the workshop is open to all. To register the workshop see ICML'06 registration procedure. It is possible to register for the workshop only (select fee waived option at
Important
Dates
Schedule
Workshop
Organizing Committee Submission
Guidelines Potential participants are invited to
submit papers according to one of the following formats: Authors should submit their
papers electronically (PDF or PS format) to the contact person (lachiche@lsiit.u-strasbg.fr)
It
is
recommended to
submit papers using the final camera-ready ICML 2006 conference paper style,
including author names. Some
relevants links
(introduction to ROC analysis, software, etc.)