Multi-view collaborative semi-supervised classification algorithm based on diversity measurers of classifier with the combination of agreement and disagreement label rules

Abstract

Considering the deficiencies of the Co-training algorithm including redundancy and sufficient conditions, diversity measurers of classifier, combined label rules, unoptimizable classifier during model updating process etc., a multi-view collaborative semi-supervised classification algorithm based on diversity measurers of the classifier with the combination of agreement and disagreement label rules (Co-AgDiag) is proposed. This algorithm uses a combination of agreement and disagreement label rules during the labelling unlabelled data process by judging whether the two classifiers are consistent and considering the diversity threshold in order to improve the performance of the classifier. During the process of generating classifier and model updating, we employed the diversity rule of classifier to update the model in order to judge the performance of the two classifiers. The experimental results on UCI datasets demonstrate the effectively and feasibility of the proposed algorithm for multi-view classification problems.

Keywords

Combined label rules diversity measurers of classifier multi-view collaborative semi-supervised classification algorithm

Get full access to this article

View all access options for this article.

References

Blum

Mitchell

(1998) Combining labeled and unlabeled data with co-training. Proceedings of the Eleventh Annual Conference on Computational Learning Theory. ACM, pp. 92–100.

Cui

Wang

Cui

(2014) AR-Tri-Training: Tri-Training with assistant strategy. Applied Mechanics and Materials 513–517: 1840–1844.

Goldman

Zhou

(2000) Enhancing supervised learning with unlabeled data. ICML. Citeseer, pp. 327–334.

Guz

Cuendet

Hakkani-Tür

. (2010) Multi-view semi-supervised learning for dialog act segmentation of speech. Processings of the IEEE Transactions on Audio, Speech, and Language 18: 320–329.

Hady

Schwenker

(2008) Co-training by committee: a new semi-supervised learning framework. Data Mining Workshops, 2008. IEEE International Conference on ICDMW’08. Washington, DC: IEEE, pp. 563–572.

Hahn

Ladner

Ostendorf

(2006) Agreement/disagreement classification: Exploiting unlabeled data using contrast classifiers. Proceedings of the Human Language Technology Conference of the NAACL, Companion Volume: Short Papers. Association for Computational Linguistics, pp. 53–56.

Hou

Jiao

(2009) Graph based Co-training algorithm for web page classification. Acta Electronica Sinica 37: 2173–2180.

Zhou

Z-H

(2007) Improve computer-aided diagnosis with machine learning techniques using undiagnosed samples. IEEE Transactions on Systems, Man and Cybernetics, Part A: Systems and Humans 37: 1088–1098.

Nigam

Ghani

(2000) Analyzing the effectiveness and applicability of co-training. Proceedings of the Ninth International Conference on Information and Knowledge Management. ACM, pp. 86–93.

10.

Settouti

El Habib Daho

El Amine Lazouni

. (2013) Random forest in semi-supervised learning (Co-Forest). 2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA). Washington DC: IEEE, pp. 326–329.

11.

Tang

H-L

Lin

Z-K

M-Y

(2008) An improved co-training text categorization algorithm based on diversity measures. Acta Electronica Sinica, S1.

12.

Zhou

Z-H

(2005a) Semi-supervised regression with co-training. IJCAI, pp. 908–913.

13.

Zhou

Z-H

(2005b) Tri-training: exploiting unlabeled data using three classifiers. IEEE Transactions on Knowledge and Data Engineering 17: 1529–1541.

14.

Zhou

Z-H

(2010) Semi-supervised learning by disagreement. Knowledge and Information Systems 24: 415–439.