TY - GEN
T1 - Induction from multi-label training examples in text categorization
T2 - 2007 International Conference on Artificial Intelligence, ICAI 2007
AU - Sarinnapakorn, Kanoksri
AU - Kubat, Miroslav
PY - 2007/12/1
Y1 - 2007/12/1
N2 - The main problem with algorithms for induction from multi-label training examples is that they often suffer from prohibitive computational costs, especially in text-categorization domains with thousands of features to characterize each document. One way to reduce these costs is to run a baseline induction algorithm separately for different subsets of features, obtaining a set of subclassifiers to be then combined. In this case study, we investigate three different ways to combine subclassifiers, including our own solution based on the Dempster-Shafer Theory.
AB - The main problem with algorithms for induction from multi-label training examples is that they often suffer from prohibitive computational costs, especially in text-categorization domains with thousands of features to characterize each document. One way to reduce these costs is to run a baseline induction algorithm separately for different subsets of features, obtaining a set of subclassifiers to be then combined. In this case study, we investigate three different ways to combine subclassifiers, including our own solution based on the Dempster-Shafer Theory.
KW - Dempster-Shafer Theory
KW - Multi-label examples
KW - Text categorization
UR - http://www.scopus.com/inward/record.url?scp=84866483573&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84866483573&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84866483573
SN - 9781601320254
T3 - Proceedings of the 2007 International Conference on Artificial Intelligence, ICAI 2007
SP - 351
EP - 357
BT - Proceedings of the 2007 International Conference on Artificial Intelligence, ICAI 2007
Y2 - 25 June 2007 through 28 June 2007
ER -