illinois-ml-nlp-users AT lists.cs.illinois.edu

Subject: Support for users of CCG software closed 7-27-20

List archive

[Illinois-ml-nlp-users] LBJ MultiLabelLearner and SVM

From: Raveesh Meena <raveesh AT csc.kth.se>
To: illinois-ml-nlp-users AT cs.uiuc.edu
Subject: [Illinois-ml-nlp-users] LBJ MultiLabelLearner and SVM
Date: Thu, 31 May 2012 18:46:30 +0200
List-archive: <http://lists.cs.uiuc.edu/pipermail/illinois-ml-nlp-users>
List-id: Support for users of CCG software <illinois-ml-nlp-users.cs.uiuc.edu>

Hi,

Has anyone used the SVM wrapper class for a MultiLabelLearner classification task?

If you are familiar with the MultiLabelLearner, then the lbj code looks like:

learn...
using...
with new MultiLabelLearner(new SparseAveragedPerceptron(.1, 0, 2))
end

I want to see how SVM perform on this task? In LBJ the MultiLabelLearner only takes LinearThresholdUnit learners (like SparseAveragedPerceptron) as input argument. How to acheive the same functionality for:

learn...
using...
with new SupportVectorMachine()
end

thanks and regards
Raveesh
--
Raveesh Meena
PhD / Graduate Student in CS

Department of Speech, Music and Hearing
Royal Institute of Technology (KTH)
Lindstedtsvägen 24
SE-100 44 Stockholm,
Sweden

Phone: +46-(0)-8-790 7872
Fax : +46-(0)-8-790 7854

Email: raveesh[at]csc.kth.se
http://www.speech.kth.se/~raveesh

[Illinois-ml-nlp-users] LBJ MultiLabelLearner and SVM, Raveesh Meena, 05/31/2012