TY - GEN

T1 - Fast learning fully complex-valued classifiers for real-valued classification problems

AU - Savitha, R.

AU - Suresh, S.

AU - Sundararajan, N.

AU - Kim, H. J.

PY - 2011

Y1 - 2011

N2 - In this paper, we present two fast learning neural network classifiers with a single hidden layer: the 'Phase Encoded Complex-valued Extreme Learning Machine (PE-CELM)' and the 'Bilinear Branch-cut Complex-valued Extreme Learning Machine (BB-CELM)'. The proposed classifiers use the phase encoded transformation and the bilinear transformation with a branch-cut at 2π as the activation functions in the input layer to map the real-valued features to the complex domain. The neurons in the hidden layer employ the fully complex-valued activation function of the type of a hyperbolic secant function. The parameters of the hidden layer are chosen randomly and the output weights are estimated as the minimum norm least square solution to a set of linear equations. The classification ability of these classifiers are evaluated using a set of benchmark data sets from the UCI machine learning repository. Results highlight the superior classification ability of these classifiers with least computational effort.

AB - In this paper, we present two fast learning neural network classifiers with a single hidden layer: the 'Phase Encoded Complex-valued Extreme Learning Machine (PE-CELM)' and the 'Bilinear Branch-cut Complex-valued Extreme Learning Machine (BB-CELM)'. The proposed classifiers use the phase encoded transformation and the bilinear transformation with a branch-cut at 2π as the activation functions in the input layer to map the real-valued features to the complex domain. The neurons in the hidden layer employ the fully complex-valued activation function of the type of a hyperbolic secant function. The parameters of the hidden layer are chosen randomly and the output weights are estimated as the minimum norm least square solution to a set of linear equations. The classification ability of these classifiers are evaluated using a set of benchmark data sets from the UCI machine learning repository. Results highlight the superior classification ability of these classifiers with least computational effort.

UR - http://www.scopus.com/inward/record.url?scp=79957799279&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79957799279&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-21105-8_70

DO - 10.1007/978-3-642-21105-8_70

M3 - Conference contribution

AN - SCOPUS:79957799279

SN - 9783642211041

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 602

EP - 609

BT - Advances in Neural Networks - 8th International Symposium on Neural Networks, ISNN 2011

T2 - 8th International Symposium on Neural Networks, ISNN 2011

Y2 - 29 May 2011 through 1 June 2011

ER -