Abstract
A novel approach to interpret the outputs of multi-class support vector machines is proposed in this paper. Using the geometrical interpretation of the classifying heperplane and the distance of the pattern from the hyperplane, one can calculate the posterior probability in binary classification case. This paper focuses on the probability output in multi-class phase where both the one-against-one and one-against-rest strategies are considered. Experiment on the speaker verification showed that this method has high performance.
This is a preview of subscription content, log in via an institution to check access.
References
Burges, C., 1998. A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery, 2:955–974.
Campbell, J., 1995. Testing with the YOHO CD-ROM Voice Verification Corpus, Proc. ICASSP: p.341–344.
Compbell, W. M., Assaleh, K. T., 1999. Polynomial Classifier Techniques for Speaker Verification, Proc. ICASSP: p.321–324.
Hastie, T., Tibshirani, R., 1996. Classification by pairwise coupling, Technical report, Stanford University and University of Toronto. http://www-stat.satnford.edu/trevor/papers/2class.ps.
Kwok, J. T. Y., 1999, Moderating the outputs of support vector machine classifiers. Neural Networks, IEEE Transactions on, 5:1018–1031.
Madevska-Bogdanova, A., Nikolic, D., 2000, A new approach of modifying SVM outputs, Proceedings of the IEEE-INNS-ENNS International Joint Conference. 6:395–398.
Platt, J., 1999, Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods, In: Advances in Large Margin Classifiers, MIT Press.
Scholkopf, B., Platt, J., Shawe-Taylor, J., Smola, A. J., Williamson, R. C., 1999. Estimating the support of a high-dimensional distribution, Microsoft Research Corporation Technical Report: MSR-TR-99-87.
Vapnik, V. N., 1999. An overview of statistical learning theory. Neural Networks, IEEE Transactions, 5:988–999.
Wan, V., Campbell, W. M., 2000. Support Vector Machines for Speaker Verification and Identification, Neural Networks for Signal Processing X. p. 775–784.
Author information
Authors and Affiliations
Corresponding author
Additional information
Project supported by the National High Technology Research & Development Programme (863) of China (No. 2001 AA4180) and Zhejiang Provincial Natural Science Foundation for Young Scientist (No. RC01058).
Rights and permissions
About this article
Cite this article
Xin, D., Wu, Zh. & Pan, Yh. Probability output of multi-class support vector machines. J. Zheijang Univ.-Sci. A 3, 131–134 (2002). https://doi.org/10.1631/BF03396426
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/BF03396426