Performance considerations for a combined information classification test using Dirichlet priors

被引：20

作者：

Lynch, RS ^{[1
]}

Willett, PK

机构：

[1] USN, Undersea Warfare Ctr, Newport, RI 02841 USA

[2] Univ Connecticut, Storrs, CT 06269 USA

来源：

IEEE TRANSACTIONS ON SIGNAL PROCESSING | 1999年 / 47卷 / 06期

关键词：

Bayesian; combining training and test data; discrete data; uniform prior;

D O I：

10.1109/78.765145

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

The average probability of error is used to demonstrate performance of the combined Bayes test (CBT), which assumes a uniform Dirichlet prior for the symbol probabilities of each class. Performance of the CBT is compared with an uncombined maximum likelihood (ML) based test and with the Kolmogorov-Smirnov test.

引用

页码：1711 / 1714

页数：4

共 20 条

[1]

[Anonymous], 1990, STAT PATTERN RECOGNI

[2]

Bernardo J.M., 2009, Bayesian Theory, V405

[3]

BOENDER CGE, 1987, BIOMETRIKA, V74, P849

[4] AVERAGING ENTROPY [J].