Connectionists: Call for Papers: ICANN 2006 Special Session

[Amaury Lendasse]

Call for Papers: ICANN 2006 Special Session:

 

Title: Feature selection and dimension reduction for regression

 

 

Abstract: Nowadays, many machine learning problems involved the use of a
large number of features. 

This might be the case, for example, in DNA and biomedical data analysis, in
image processing,

financial data mining, chemometrics, etc.  In other cases, the number of
features may be smaller, but

of the same order of magnitude as the number of samples.  In both cases,
regression tasks are

 faced to the curse of dimensionality: overfitting easily appears, and in
some cases the regression

problem can become ill-posed (or not identifiable).  The challenge is then
to reduce the number of

features, in order to improve the regression efficiency.  Interpretability
is often a major concern too,

as a large number of features usually prevents any understanding of the
underlying relationship.

 

Feature selection and dimension reduction includes two different ways of
reducing the number

inputs of the regression model.  First, inputs can be selected among the
original features; this is

usually referred to as feature selection or input selection.  Second, inputs
can be built from the

original features, by combining them in a linear or nonlinear way; this
leads to dimension reduction

(sometimes referred to as variable selection).

 

The goal of feature selection and dimension reduction is twofold.  First,
reducing the number of

input variables fights the curse of dimensionality, giving the possibility
of increasing the regression

generalization performances.  Second, a reduced set of variables is of
utmost importance in real

applications as it allows an easier interpretation of the relationship
between features and outputs.

 

The aim of this session is to present original developments in feature
selection and dimension reduction. 

Contributions are invited in the following areas:

- new algorithms and methods;

- comparisons between techniques, including the assessment of the compromise
between

generalization properties and computational load;

-applicability of the proposed methods in real-world problems, including
small sample and high

dimension constraints.

 

It is suggested (but not mandatory) to illustrate and compare the proposed
methods by using one of

or both the following regression datasets: Housing (Boston), available from
the UCI Machine

Learning Repository (http://www.ics.uci.edu/~mlearn/MLSummary.html), and
Orange juice spectra, available

from the UCL Machine Learning Group website
(http://www.ucl.ac.be/mlg/index.php?page=DataBases). 

 

Practical details:

- ICANN’06 website: http://icann2006.org/chapter1/index.html

- 30 Mars: End of submission of papers to special sessions

- Proceedings of ICANN will be published in Springer's "Lecture Notes in
Computer Science" series.

Paper length is restricted to a maximum of 10 pages, including figures.

 

Organized by:

Amaury Lendasse, Helsinki University of Technology, Adaptive Informatics
Research Centre, Finland.

Michel Verleysen, Université catholique de Louvain, Machine Learning group,
Belgium.