Abstract
Cultural modeling aims at developing behavioral models of groups and analyzing the impact of culture factors on group behavior using computational methods. Machine learning methods in particular classification, play a central role in such applications. In modeling cultural data, it is expected that standard classifiers yield good performance under the assumptions that class distribution is balanced and different classification errors have uniform costs. However, these assumptions are often violated in practice and thus the performance of standard classifiers is severely hindered. To handle this problem, this paper studies cost-sensitive learning in cultural modeling domain by considering cost factor when building the classifiers, with the aim of minimizing total misclassification costs. We empirically investigate four typical cost-sensitive learning methods, combine them with six standard classifiers and evaluate their performances under various conditions. Our empirical study verifies the effectiveness of cost-sensitive learning in cultural modeling. Based on the results of our experimental study, we gain a thorough insight into the problem of class imbalance and non-uniform misclassification costs, as well as the selection of cost-sensitive methods, base classifiers and method-classifier pairs for this domain.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings of 2010 IEEE International Conference on Service Operations and Logistics, and Informatics, SOLI 2010 |
| Pages | 398-403 |
| Number of pages | 6 |
| DOIs | |
| State | Published - 2010 |
| Event | 2010 IEEE International Conference on Service Operations and Logistics, and Informatics, SOLI 2010 - QingDao, China Duration: Jul 15 2010 → Jul 17 2010 |
Other
| Other | 2010 IEEE International Conference on Service Operations and Logistics, and Informatics, SOLI 2010 |
|---|---|
| Country | China |
| City | QingDao |
| Period | 7/15/10 → 7/17/10 |
Fingerprint
Keywords
- Class imbalance problem
- Classification
- Cost-sensitive learning
- Cultural modeling
ASJC Scopus subject areas
- Information Systems and Management
- Management Science and Operations Research
Cite this
Employing cost-sensitive learning in cultural modeling. / Su, Peng; Mao, Wenji; Zeng, Dajun; Wang, Fei Yue.
Proceedings of 2010 IEEE International Conference on Service Operations and Logistics, and Informatics, SOLI 2010. 2010. p. 398-403 5551546.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Employing cost-sensitive learning in cultural modeling
AU - Su, Peng
AU - Mao, Wenji
AU - Zeng, Dajun
AU - Wang, Fei Yue
PY - 2010
Y1 - 2010
N2 - Cultural modeling aims at developing behavioral models of groups and analyzing the impact of culture factors on group behavior using computational methods. Machine learning methods in particular classification, play a central role in such applications. In modeling cultural data, it is expected that standard classifiers yield good performance under the assumptions that class distribution is balanced and different classification errors have uniform costs. However, these assumptions are often violated in practice and thus the performance of standard classifiers is severely hindered. To handle this problem, this paper studies cost-sensitive learning in cultural modeling domain by considering cost factor when building the classifiers, with the aim of minimizing total misclassification costs. We empirically investigate four typical cost-sensitive learning methods, combine them with six standard classifiers and evaluate their performances under various conditions. Our empirical study verifies the effectiveness of cost-sensitive learning in cultural modeling. Based on the results of our experimental study, we gain a thorough insight into the problem of class imbalance and non-uniform misclassification costs, as well as the selection of cost-sensitive methods, base classifiers and method-classifier pairs for this domain.
AB - Cultural modeling aims at developing behavioral models of groups and analyzing the impact of culture factors on group behavior using computational methods. Machine learning methods in particular classification, play a central role in such applications. In modeling cultural data, it is expected that standard classifiers yield good performance under the assumptions that class distribution is balanced and different classification errors have uniform costs. However, these assumptions are often violated in practice and thus the performance of standard classifiers is severely hindered. To handle this problem, this paper studies cost-sensitive learning in cultural modeling domain by considering cost factor when building the classifiers, with the aim of minimizing total misclassification costs. We empirically investigate four typical cost-sensitive learning methods, combine them with six standard classifiers and evaluate their performances under various conditions. Our empirical study verifies the effectiveness of cost-sensitive learning in cultural modeling. Based on the results of our experimental study, we gain a thorough insight into the problem of class imbalance and non-uniform misclassification costs, as well as the selection of cost-sensitive methods, base classifiers and method-classifier pairs for this domain.
KW - Class imbalance problem
KW - Classification
KW - Cost-sensitive learning
KW - Cultural modeling
UR - http://www.scopus.com/inward/record.url?scp=77957821772&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77957821772&partnerID=8YFLogxK
U2 - 10.1109/SOLI.2010.5551546
DO - 10.1109/SOLI.2010.5551546
M3 - Conference contribution
AN - SCOPUS:77957821772
SN - 9781424471188
SP - 398
EP - 403
BT - Proceedings of 2010 IEEE International Conference on Service Operations and Logistics, and Informatics, SOLI 2010
ER -