(403) previous ~ index ~ next
To: tdt-distrib@unagi.cis.upenn.edu
From: David Graff <graff@unagi.cis.upenn.edu>
Subject: Re: any tdt svm experiments [from Ralf Klinkenberg]
Date: Mon, 21 Oct 2002 09:47:26 -0400
------- Forwarded Message
Date: Mon, 21 Oct 2002 10:43:11 +0200
From: Ralf Klinkenberg <klinkenberg@kimo.cs.uni-dortmund.de>
To: Christine.Piatko@jhuapl.edu
CC: David Graff <graff@unagi.cis.upenn.edu>, tdt-distrib@unagi.cis.upenn.edu
Subject: Re: any tdt svm experiments [from Christine.Piatko@jhuapl.edu]
Thorsten Joachims and I presented a framework for SVMs for learning
drifting concepts at ICML-2000:
@InProceedings{Klinkenberg/Joachims/2000a,
author = {Klinkenberg, Ralf and Joachims, Thorsten},
title = {Detecting Concept Drift with Support Vector Machines},
booktitle = {Proceedings of the Seventeenth International Conference on Machine Learning (ICML)},
publisher = {Morgan Kaufmann},
address = {San Francisco, CA, USA},
editor = {Langley, Pat},
year = {2000},
pages = {487--494},
url = {http://www-ai.cs.uni-dortmund.de/DOKUMENTE/klinkenberg_joachims_2000a.ps.gz}
}
We have not applied this approach to any TDT data yet, but I guess
it might be applicable to the TDT tracking task as well, since
learning a concept that changes over time and keeping the learned
hypothesis up to date seems to be quite similar to tracking a
concept. But as I said, that's a just a guess and we haven't done
any experiments with the TDT data.
Best regards,
Ralf
+-----------------------------+--------------------------------------------+
| Ralf Klinkenberg | klinkenberg@ls8.cs.uni-dortmund.de ___ |
| LS Künstliche Intelligenz | http://www-ai.cs.uni-dortmund.de //// |
| Fachbereich Informatik VIII | UNI DO// |
| Universität Dortmund | Tel.: (+49 231) 755-5103 ___ //// |
| D - 44221 Dortmund | Fax: (+49 231) 755-5105 \\\\/// |
| Germany | Büro: GB IV / R 117 \\\\/ |
+-----------------------------+--------------------------------------------+
------- End of Forwarded Message
-------------------------------------------------------------
To unsubscribe from tdt-distrib, email majordomo@ldc.upenn.edu
with "unsubscribe tdt-distrib" in the body of the message.
(403) previous ~ index ~ next
Last updated Mon Nov 11 14:16:28 2002