Reminder: Fall Seminar on Data and Knowledge Integration (this week)

[Chitta Baral <chitta@asu.edu>]

The eighth talk of the seminar series
  will be this Friday.
 
   Date: 10/25/02 (Friday)
   Time: 11:00 -- 12:15 PM
   Place: GWC 487
  
  Speaker: Professor Huan Liu.

  Title: From Source Description to Feature Selection

  (Based on an ICML'02 paper titled
   ``Feature selection with selective sampling''
   by: Huan Liu, Hiroshi Motoda and Lei Yu available at
   http://www.public.asu.edu/~huanliu/papers/icml02.ps)

The talk introduces a technique - feature selection that can be applied
to source description. It's based on our work published at this year's
International Conference on Machine Learning. Feature selection, as a
preprocessing step to machine learning, has been shown very effective in
reducing dimensionality, removing irrelevant data, increasing learning
curacy, and improving comprehensibility. In particular, we consider
the problem of active feature selection in a filter model setting. We
describe a formalism of active feature selection called selective
sampling, demonstrate it by applying it to a widely used feature
selection algorithm Relief, and show how it realizes active
feature selection and reduces the required number of training data
for Relief to achieve time savings without performance deterioration. We
will explore the usage of feature selection in source description.

----

The tentative schedule for November is as follows:

Nov 1: Prof. Goran Konjevod 
Nov 8th: Prof. Rao Kambhampati
Nov 15th: Prof. Selcuk Candan 
Nov 22nd: Prof. Huan Liu
Nov 29th: Thanksgiving weekend

-- 

-- 
 Chitta Baral
 Dept. of Computer Sc. and Engg.,
 Arizona State University, Tempe, AZ 85287, USA.
 chitta@asu.edu, http://www.public.asu.edu/~cbaral/
 Ph: 480-727-6047, Fax: 480-965-2751