A radial basis function network classifier to maximise leave-one-out mutual information

Download

Full text not archived in this repository.

Advice

Please see our End User Agreement.

It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.

Tools

Lists

Hong, X. ORCID: https://orcid.org/0000-0002-6832-2298, Chen, S., Qatawneh, A., Daqrouq, K., Sheikh, M. and Morfeq, A. (2014) A radial basis function network classifier to maximise leave-one-out mutual information. Applied Soft Computing, 23. pp. 9-18. ISSN 1568-4946 doi: 10.1016/j.asoc.2014.06.003

Abstract/Summary

tWe develop an orthogonal forward selection (OFS) approach to construct radial basis function (RBF)network classifiers for two-class problems. Our approach integrates several concepts in probabilisticmodelling, including cross validation, mutual information and Bayesian hyperparameter fitting. At eachstage of the OFS procedure, one model term is selected by maximising the leave-one-out mutual infor-mation (LOOMI) between the classifier’s predicted class labels and the true class labels. We derive theformula of LOOMI within the OFS framework so that the LOOMI can be evaluated efficiently for modelterm selection. Furthermore, a Bayesian procedure of hyperparameter fitting is also integrated into theeach stage of the OFS to infer the l2-norm based local regularisation parameter from the data. Since eachforward stage is effectively fitting of a one-variable model, this task is very fast. The classifier construc-tion procedure is automatically terminated without the need of using additional stopping criterion toyield very sparse RBF classifiers with excellent classification generalisation performance, which is par-ticular useful for the noisy data sets with highly overlapping class distribution. A number of benchmarkexamples are employed to demonstrate the effectiveness of our proposed approach.

Altmetric Badge

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/37201
Identification Number/DOI	10.1016/j.asoc.2014.06.003
Refereed	Yes
Divisions	Science > School of Mathematical, Physical and Computational Sciences > Department of Computer Science
Uncontrolled Keywords	Cross validationMutual informationOrthogonal forward selectionRadial basis function classifiera
Publisher	Elsevier
Download/View statistics	View download statistics for this item

Deposit Details

Date Deposited:	31 Jul 2014 07:15	Date item deposited into CentAUR
Last Modified:	23 Jun 2024 04:52	Date item last modified

References

References[1] M. Stone, Cross-validatory choice and assessment of statistical predictions, J.R. Stat. Soc. Ser. B 36 (2) (1974) 117–147.[2] R.H. Myers, Classical and Modern Regression with Applications, 2nd ed., PWS-KENT, Boston, 1990.[3] S. Chen, S.A. Billings, W. Luo, Orthogonal least squares methods and theirapplications to non-linear system identification, Int. J. Control 50 (5) (1989)1873–1896.[4] M.J. Korenberg, Identifying nonlinear difference equation and functionalexpansion representations: the fast orthogonal algorithm, Ann. Biomed. Eng.16 (1) (1988) 123–142.[5] S. Chen, C.F.N. Cowan, P.M. Grant, Orthogonal least squares learning algo-rithm for radial basis function networks, IEEE Trans. Neural Netw. 2 (2) (1991)302–309.[6] L.-X. Wang, J.M. Mendel, Fuzzy basis functions, universal approximation, andorthogonal least-squares learning, IEEE Trans. Neural Netw. 5 (5) (1992)807–814.[7] X. Hong, C.J. Harris, Neurofuzzy design and model construction of nonlineardynamical processes from data, IEE Proc. Control Theory Appl. 148 (6) (2001)530–538.[8] Q. Zhang, Using wavelets network in nonparametric estimation, IEEE Trans.Neural Netw. 8 (2) (1997) 227–236.[9] S.A. Billings, H.L. Wei, The wavelet-NARMAX representation: a hybrid modelstructure combining polynomial models with multiresolution wavelet decom-positions, Int. J. Syst. Sci. 36 (3) (2005) 137–152.[10] N. Chiras, C. Evans, D. Rees, Nonlinear gas turbine modeling using NARMAXstructures, IEEE Trans. Instrum. Meas. 50 (4) (2001) 893–898.[11] Y. Gao, M.J. Er, Online adaptive fuzzy neural identification and control ofa class of MIMO nonlinear systems, IEEE Trans. Fuzzy Syst. 11 (4) (2003)462–477.[12] K.M. Tsang, W.L. Chan, Adaptive control of power factor correction converterusing nonlinear system identification, IEE Proc. Electr. Power Appl. 152 (3)(2005) 627–633.[13] G.-C. Luh, W.-C. Cheng, Identification of immune models for fault detection,Proc. Inst. Mech. Eng. I: J. Syst. Control Eng. 218 (2004) 353–367. [14] G.W. Chang, C. Chen, Y. Liu, A neural-network-based method of modeling elec-tric arc furnace load for power engineering study, IEEE Trans. Power Syst. 25(1) (2010) 138–146.[15] B. Mutnury, M. Swaminathan, J.P. Libous, Macromodeling of nonlinear digitalI/O drivers, IEEE Trans. Adv. Pack. 29 (1) (2006) 102–113.[16] C. Huang, F. Wang, An RBF network with OLS and EPSO algorithms for real-timepower dispatch, IEEE Trans. Power Syst. 22 (1) (2007) 96–104.[17] R. Mukai, V.A. Vilnrotter, P. Arabshahi, V. Jammejad, Adaptive acquisition andtracking for deep space array feed antennas, IEEE Trans. Neural Netw. 13 (5)(2002) 1149–1162.[18] V.S. Kodogiannis, J.N. Lygouras, A. Tarczynski, H.S. Chowdrey, Artificial odordiscrimination system using electronic nose and neural networks for the iden-tification of urinary tract infection, IEEE Trans. Inf. Technol. Biomed. 12 (6)(2008) 707–713.[19] C. Kauffmann, P. Motreff, L. Sarry, In vivo supervised analysis of stent reen-dothelialization from optical coherence tomography, IEEE Trans. Med. Imaging29 (3) (2010) 807–818.[20] G.P. Asner, R.E. Martin, R. Tupayachi, et al., Taxonomy and remote sensing ofleaf mass per area (LMA) in humid tropical forests, Ecol. Appl. 21 (1) (2011)85–98.[21] M.J.L. Orr, Regularization in the selection of radial basis function centers, NeuralComput. 7 (3) (1995) 606–623.[22] S. Chen, E.S. Chng, K. Alkadhimi, Regularised orthogonal least squares algorithmfor constructing radial basis function networks, Int. J. Control 64 (5) (1996)829–837.[23] S. Chen, Y. Wu, B.L. Luk, Combined genetic algorithm optimization and regular-ized orthogonal least squares learning for radial basis function networks, IEEETrans. Neural Netw. 10 (5) (1999) 1239–1243.[24] M.E. Tipping, Sparse Bayesian learning and the relevance vector machine, J.Mach. Learn. Res. 1 (2001) 211–244.[25] S. Chen, X. Hong, C.J. Harris, Sparse kernel regression modelling using combinedlocally regularised orthogonal least squares and D-optimality experimentaldesign, IEEE Trans. Autom. Control 48 (6) (2003) 1029–1036. [26] S. Chen, X. Hong, C.J. Harris, P.M. Sharkey, Sparse modelling using forwardregression with PRESS statistic and regularization, IEEE Trans. Syst. Man Cybern.B 34 (2) (2004) 898–911.[27] S. Chen, Local regularization assisted orthogonal least squares regression, Neu-rocomputing 69 (4-6) (2006) 559–585.[28] D.J.C. MacKay, Bayesian Methods for Adaptive Models (Ph.D. Thesis), CaliforniaInstitute of Technology, USA, 1992.[29] C.E. Shannon, A mathematical theory of information, Bell Syst. Tech. J. 27 (1948)379–423.[30] T.M. Cover, J.A. Thomas, Elements of Information Theory, John Wiley & Sons,1991.[31] G.L. Zheng, S.A. Billings, Radial basis function network configuration usingmutual information and the orthogonal least squares algorithm, Neural Netw.9 (9) (1996) 1619–1637.[32] F. Rossi, A. Lendasse, D. Franc¸ ois, V. Wertz, M. Verleysen, Mutual informationfor the selection of relevant variables in spectrometric nonlinear modelling,Chemom. Intell. Lab. Syst. 80 (2) (2006) 215–226.[33] H. Peng, F. Long, C. Ding, Feature selection based on mutual information criteriaof max-dependency, max-relevance and min-redundancy, IEEE Trans. PatternAnal. Mach. Intell. 27 (8) (2005) 1226–1238.[34] J.P.W. Pluim, J.B.A. Maintz, M.A. Viergever, Mutual information based regis-tration of medical images: a survey, IEEE Trans. Med. Imaging 22 (8) (2003)986–1004.[35] X. Zhou, X. Wang, E.R. Dougherty, Nonlinear probit gene classification usingmutual information and wavelet-based feature selection, J. Biol. Syst. 12 (3)(2004) 371–386.[36] X. Hong, S. Chen, C.J. Harris, A fast kernel classifier construction algorithm usingorthogonal forward selection to minimize leave-one-out misclassification rate,Int. J. Syst. Sci. 39 (2) (2008) 119–125.[37] X. Hong, S. Chen, C.J. Harris, A kernel-based two class classifier for imbalanceddata sets, IEEE Trans. Neural Netw. 18 (1) (2007) 28–41.[38] X. Hong, P.M. Sharkey, K. Warwick, Automatic nonlinear predictive model con-struction algorithm using forward regression and the PRESS statistic, IEE Proc.Control Theory Appl. 150 (3) (2003) 245–254.[39] B.D. Ripley, Pattern Recognition and Neural Networks, Cambridge UniversityPress, Cambridge, 1996.[40] G. Rätsch, T. Onoda, K.R. Müller, Soft margins for AdaBoost, Mach. Learn. 42 (3)(2001) 287–320.[41] G. Rätsch, http://www.raetschlab.org/members/raetsch

University Staff: Request a correction | Centaur Editors: Update this record

Search Google Scholar