A generalization of the minimum classification error (MCE) training method for speech recognition and detection

A generalization of the minimum classification error (MCE) training method for speech recognition and detection

The model training algorithm is a critical component in the statistical pattern recognition approaches which are based on the Bayes decision theory. Conventional applications of the Bayes decision theory usually assume uniform error cost and result in a ubiquitous use of the maximum a posteriori (MA...

Full description

Bibliographic Details
Main Author: Fu, Qiang
Published: Georgia Institute of Technology 2008
Subjects:
MVE
P-MCE
Non-uniform error cost
Automatic speech recognition
Bayesian statistical decision theory
Pattern recognition systems
Online Access:http://hdl.handle.net/1853/22705

Internet

http://hdl.handle.net/1853/22705

Similar Items