Semi-Supervised Least-Squares Support Vector Classifier based on Virtual Leave One Out Residuals

Stanislaw Jankowski, Zbigniew Szymański, Ewa Piatkowska-Janko

Abstract

We present a new semi-supervised learning system based on least-squares support vector machine classifier. We apply the virtual leave-one-out residuals as criterion for selection of the most influential data for label switching test. The analytic form of the solution enables to obtain a high gain of the computational cost. The quality of the method was tested on the artificial data set – two moons problem and on the real signal-averaged ECG data set. The correct classification score is better as compared to other methods.

References

  1. 1. V.N. Vapnik: Statistical Learning Theory, Wiley Interscience, New York 1998.
  2. O. Chapelle, B. Schölkopf, A. Zien: Semi-supervised Learning, MIT Press, Cambridge 2006.
  3. J.A.K. Suykens and J. Vandewalle: Least squares support vector machine classifier, Neural Processing Letters, 9 (1999), 293-300.
  4. J.A.K. Suykens, T. Van Gestel, J. De Brabanter, B. De Moor, J. Vandewalle: Least Squares Support Vector Machines, World Scientific 2002.
  5. R.D. Cook, S. Weisberg, Characterization of an empirical influence function for detecting influential cases in regression, Technometrics 22 (1980), 495-508.
  6. G. Monari, G. Dreyfus: Local Overfitting Control via Leverages; Neural Computation 14 (2002), 1481-1506.
  7. G.C. Cawley, N.L.C. Talbot: Fast exact leave-one-out cross-validation of sparse leastsquares support vector machines, Neural Networks 17 (2004), 1467-1475.
  8. G.C. Cawley, N.L.C. Talbot: Preventing Over-fitting during Model Selection via Bayesian Regularisation of the Hyper-Parameters, Journal of Machine Learning Research 8, (2007), 841-861.
  9. F. Ojeda, J. A.K. Suykens, B. De Moor: Low rank updated LS-SVM classifier for fast variable selection, Neural Networks 21 (2008), 437-449.
  10. K. Bennett, A. Demiriz: Semi-supervised support vector machines, in M.S. Kearns, S.A. Solla and D.A. Cohn (eds.) Advances in Neural Information Processing Systems 12, MIT Press, Cambridge, MA, 1998, 368-374.
  11. O. Chapelle, V. Sindwani, S.A. Keerthi: Optimization Techniques for Semi-Supervised Support Vector Machines, Journal of Machine Learning Research 9, 2008, 203-233.
  12. J.A. Gomes: Signal averaged electrocardiography - concepts, methods and applications. Kluwer Academic Publishers, 1993.
  13. S. Jankowski, Z. Szymanski, E. Piatkowska-Janko, A. Oreziak: Improved recognition of sustained ventricular tachycardia from SAECG by support vector machine, Anatolian Journal of Cardiology, vol. 7 (2007), 112-115.
  14. S. Jankowski, E. Piatkowska-Janko, Z. Szymanski and A. Oreziak: Transductive Support Vector Machines for Risk Recognition of Sustained Ventricular Tachycardia and Flicker after Myocardial Infarction, in Proceedings of the 7th International Workshop on Pattern Recognition in Information Systems - PRIS 2007 (eds. Ana Fred, Anil K. Jain), June 2007, Funchal, Portugal, 161-170.
Download


Paper Citation


in Harvard Style

Jankowski S., Szymański Z. and Piatkowska-Janko E. (2009). Semi-Supervised Least-Squares Support Vector Classifier based on Virtual Leave One Out Residuals . In Proceedings of the 9th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2009) ISBN 978-989-8111-89-0, pages 71-80. DOI: 10.5220/0002196400710080


in Bibtex Style

@conference{pris09,
author={Stanislaw Jankowski and Zbigniew Szymański and Ewa Piatkowska-Janko},
title={Semi-Supervised Least-Squares Support Vector Classifier based on Virtual Leave One Out Residuals},
booktitle={Proceedings of the 9th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2009)},
year={2009},
pages={71-80},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002196400710080},
isbn={978-989-8111-89-0},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 9th International Workshop on Pattern Recognition in Information Systems - Volume 1: PRIS, (ICEIS 2009)
TI - Semi-Supervised Least-Squares Support Vector Classifier based on Virtual Leave One Out Residuals
SN - 978-989-8111-89-0
AU - Jankowski S.
AU - Szymański Z.
AU - Piatkowska-Janko E.
PY - 2009
SP - 71
EP - 80
DO - 10.5220/0002196400710080