TY - GEN

T1 - A new algorithm for learning mahalanobis discriminant functions by a neural network

AU - Ito, Yoshifusa

AU - Izumi, Hiroyuki

AU - Srinivasan, Cidambi

PY - 2011

Y1 - 2011

N2 - It is well known that a neural network can learn Bayesian discriminant functions. In the two-category normal-distribution case, a shift by a constant of the logit transform of the network output approximates a corresponding Mahalanobis discriminant function [7]. In [10], we have proposed an algorithm for estimating the constant, but it requires the network to be trained twice, in one of which the teacher signals must be shifted by the mean vectors. In this paper, we propose a more efficient algorithm for estimating the constant with which the network is trained only once.

AB - It is well known that a neural network can learn Bayesian discriminant functions. In the two-category normal-distribution case, a shift by a constant of the logit transform of the network output approximates a corresponding Mahalanobis discriminant function [7]. In [10], we have proposed an algorithm for estimating the constant, but it requires the network to be trained twice, in one of which the teacher signals must be shifted by the mean vectors. In this paper, we propose a more efficient algorithm for estimating the constant with which the network is trained only once.

UR - http://www.scopus.com/inward/record.url?scp=81855218298&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=81855218298&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-24958-7_69

DO - 10.1007/978-3-642-24958-7_69

M3 - Conference contribution

AN - SCOPUS:81855218298

SN - 9783642249570

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

SP - 596

EP - 605

BT - Neural Information Processing - 18th International Conference, ICONIP 2011, Proceedings

T2 - 18th International Conference on Neural Information Processing, ICONIP 2011

Y2 - 13 November 2011 through 17 November 2011

ER -