Journal of Software, Vol 7, No 6 (2012), 1307-1314, Jun 2012
doi:10.4304/jsw.7.6.1307-1314

Confidence Estimation for Graph-based Semi-supervised Learning

Tao Guo, Guiyang Li

Abstract


To select unlabeled example effectively and reduce classification error, confidence estimation for graph-based semi-supervised learning （CEGSL） is proposed. This algorithm combines graph-based semi-supervised learning with collaboration-training. It makes use of structure information of sample to calculate the classification probability of unlabeled example explicitly. With multi-classifiers, the algorithm computes the confidence of unlabeled example implicitly. With dual-confidence estimation, the unlabeled example is selected to update classifiers. The comparative experiments on UCI datasets indicate that CEGSL can effectively exploit unlabeled data to enhance the learning performance.


Keywords


graph, collaboration-training, confidence, classification, semi-supervised leaning,

References


 

[1] I. Cohen, F. G. Cozman, N. Sebe, M. C. Cirelo, T. S. Huang. Semisupervised learning of classifiers: Theory, algorithm, and their application to human-computer interaction[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(12): 1553-1567.
http://dx.doi.org/10.1109/TPAMI.2004.127
PMid:15573817

[2] Z.-H. Zhou. Learning with unlabeled data and its application to image retrieval. In Proceedings of the 9th Pacific Rim International Conference on Artificial Intelligence, pages 5–10(2006).

[3] Jimin Li, A Novel Semi-supervised SVM Based on Tri-training for Intrusion Detection. JOURNAL OF COMPUTERS, VOL. 5, NO. 4, APRIL (2010).

[4] Kurt Driessens1, Peter Reutemann2, Using Weighted Nearest Neighbor to Benefit from Unlabeled Data. Encyclopedia of Machine Learning 2010: 857-862(2010).

[5] A. Blum, T. Mitchell. Combining labeled and unlabeled example with co-training[C]. In: Proceedings of the 11th Annual Conference on Computational Learning Theory (COLT’98), Wisconsin, MI, 1998, pp:92-100.

[6] Zhu X. Semi-supervised learning literature survey [R]. Technical Report 1530, Department of Computer Sciences, University of Wisconsin at Madison, Madison, WI, Jul.2008.

[7] D. Zhou, O. Bousquet, T. Lal etc. Learning with local and global consistency [C]. Advances in Neural Information Processing Systems(NIPS), Cambridge, MA:MIT Press, 2004, 16, pp:321- 328.

[8] S. Goldman, Y. Zhou. Enhancing supervised learning with unlabeled example[C]. In: Proceedings of the 17th International Conference on Machine Learning (ICML’00), San Francisco, CA, 2000, pp:327-334.

[9] X.J. Zhu. Z. Ghahramani. Semi-supervised learning using Gaussian fields and Harmonic functions[C]. In: Proceedings of International Conference of Machine Learning(ICML’03), Washington DC, 2003, pp:912-919.

[10] Zhou Z H, Li M. Tri-training: Exploiting unlabeled example using three classifiers [J].IEEE Transactions on Knowledge and Data Engineering,2005,17(11), pp:1529-1541.
http://dx.doi.org/10.1109/TKDE.2005.186

[11] Tao Guo, Guiyang Li, Improved Tri-Training with Unlabeled example [C]. International Conference on Nanotechnology and Computer Engineering (CNCE), 2011.

[12] Yongzhao Zhan,Yabi Cheng. Co-Training Semi-Supervised Active Learning Algorithm with Noise Filter[J]. Pattern Recognition and Artificial Intelligence, 2009,22(5):750-755

[13] Zhu X J. Semi-supervised learning with graphs [D].USA: Carnegie Mellon University. 2006.1-89.

[14] Blum A, Chawla S. Learning from labeled and unlabeled example using graph mincuts [C]. in: Proceeding of 18th International Conference on Machine Learning. 2001.

[15] Pavan K, Rong J. SemiBoost: Boost for Semi-supervised Learning [J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(11):2000-2014.
http://dx.doi.org/10.1109/TPAMI.2008.235
PMid:19762927

[16] Zhou Z H, Wang Jue, Machine learning and application [M], Beijing, Tsinghua University Press, 2007, PP:259-275


Full Text: PDF


Journal of Software (JSW, ISSN 1796-217X)

Copyright @ 2006-2013 by ACADEMY PUBLISHER – All rights reserved.