TY - CHAP

T1 - Multicategory Bayesian decision using a three-layer neural network

AU - Ito, Yoshifusa

AU - Srinivasan, Cidambi

PY - 2003

Y1 - 2003

N2 - We realize a multicategory Bayesian classifier by a three-layer neural network having rather a small number of hidden layer units. The state-conditional probability distributions are supposed to be multivariate normal distributions. The network has direct connections between the input and output layers. Its outputs are monotone mappings of posterior probabilities. Hence, they can be used as discriminant functions and, in addition, the posterior probabilities can be easily retrieved from the outputs.

AB - We realize a multicategory Bayesian classifier by a three-layer neural network having rather a small number of hidden layer units. The state-conditional probability distributions are supposed to be multivariate normal distributions. The network has direct connections between the input and output layers. Its outputs are monotone mappings of posterior probabilities. Hence, they can be used as discriminant functions and, in addition, the posterior probabilities can be easily retrieved from the outputs.

UR - http://www.scopus.com/inward/record.url?scp=35248827534&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35248827534&partnerID=8YFLogxK

U2 - 10.1007/3-540-44989-2_31

DO - 10.1007/3-540-44989-2_31

M3 - Chapter

AN - SCOPUS:35248827534

SN - 3540404082

SN - 9783540404088

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

SP - 253

EP - 261

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

A2 - Kaynak, Okyay

A2 - Alpaydin, Ethem

A2 - Oja, Erkki

A2 - Xu, Lei

ER -