Deterministic neural classification.

by: KA Toh

Neural computation, Vol. 20, No. 6. (June 2008), pp. 1565-1595.

View FullText article

DOI, Pubmed, Hubmed

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

sen_cheng, bayesian

Find related articles with these CiteULike tags

cluster, coding, theory

Abstract

This letter presents a minimum classification error learning formulation for a single-layer feedforward network (SLFN). By approximating the nonlinear counting step function using a quadratic function, the classification error rate is shown to be deterministically solvable. Essentially the derived solution is related to an existing weighted least-squares method with class-specific weights set according to the size of data set. By considering the class-specific weights as adjustable parameters, the learning formulation extends the classification robustness of the SLFN without sacrificing its intrinsic advantage of being a closed-form algorithm. While the method is applicable to other linear formulations, our empirical results indicate SLFN's effectiveness on classification generalization.

@article{citeulike:2783163, abstract = {This letter presents a minimum classification error learning formulation for a single-layer feedforward network (SLFN). By approximating the nonlinear counting step function using a quadratic function, the classification error rate is shown to be deterministically solvable. Essentially the derived solution is related to an existing weighted least-squares method with class-specific weights set according to the size of data set. By considering the class-specific weights as adjustable parameters, the learning formulation extends the classification robustness of the SLFN without sacrificing its intrinsic advantage of being a closed-form algorithm. While the method is applicable to other linear formulations, our empirical results indicate SLFN's effectiveness on classification generalization.}, address = {Biometrics Engineering Research Center, School of Electrical and Electronic Engineering, Yonsei University, Seoul 120-749, Korea, katoh@yonsei.ac.kr.}, author = {Toh, K. A. }, citeulike-article-id = {2783163}, doi = {10.1162/neco.2007.04-07-508}, issn = {0899-7667}, journal = {Neural computation}, keywords = {cluster, coding}, month = {June}, number = {6}, pages = {1565--1595}, posted-at = {2008-05-10 20:05:19}, priority = {2}, title = {Deterministic neural classification.}, url = {http://dx.doi.org/10.1162/neco.2007.04-07-508}, volume = {20}, year = {2008} }

RIS record

TY - JOUR ID - citeulike:2783163 L3 - citeulike-article-id:2783163 TI - Deterministic neural classification. JF - Neural computation VL - 20 IS - 6 SP - 1565 EP - 1595 AD - Biometrics Engineering Research Center, School of Electrical and Electronic Engineering, Yonsei University, Seoul 120-749, Korea, katoh@yonsei.ac.kr. SN - 0899-7667 N2 - This letter presents a minimum classification error learning formulation for a single-layer feedforward network (SLFN). By approximating the nonlinear counting step function using a quadratic function, the classification error rate is shown to be deterministically solvable. Essentially the derived solution is related to an existing weighted least-squares method with class-specific weights set according to the size of data set. By considering the class-specific weights as adjustable parameters, the learning formulation extends the classification robustness of the SLFN without sacrificing its intrinsic advantage of being a closed-form algorithm. While the method is applicable to other linear formulations, our empirical results indicate SLFN's effectiveness on classification generalization. KW - cluster KW - coding AU - Toh, KA PY - 2008/06// UR - http://dx.doi.org/10.1162/neco.2007.04-07-508 ER -

RIS BibTeX