Estimating cell frequencies under inequality constraints based on the Kullback-Leibler information

This article considers the problem of estimating the cell frequencies in a contingency table under inequality constraints. Algorithms are proposed for cell frequency estimation via minimizing the Kullback–Leibler distance subject to inequality constraints. The proposed algorithms are shown to be sim...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Statistica Neerlandica Jg. 66; H. 2; S. 203 - 216
Hauptverfasser: Fu, Lianyan, Gao, Wei, Shi, Ning-Zhong
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Oxford, UK Blackwell Publishing Ltd 01.05.2012
Netherlands Society for Statistics and Operations Research
Schriftenreihe:Statistica Neerlandica
Schlagworte:
Algorithms
Contingency
Convergence
Estimating
Estimating techniques
Frequencies
I-projection
Inequalities
inequality constraints
maximum entropy
optimization
Statistical analysis
Studies
Tables (data)
the minimum discrimination information
Theorems
ISSN:0039-0402, 1467-9574
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This article considers the problem of estimating the cell frequencies in a contingency table under inequality constraints. Algorithms are proposed for cell frequency estimation via minimizing the Kullback–Leibler distance subject to inequality constraints. The proposed algorithms are shown to be simple, easy to be used, fast, and reliable. Theorems are derived to guarantee the convergence of the algorithms. Applications and extensions of the algorithms are provided for more general problems than contingency table. The R programs that implement the proposed algorithms are presented in Appendix B.
Bibliographie:istex:9C3DC8EEA7404F49F4C823626B2FB8DA45D8FD00
ArticleID:STAN513
ark:/67375/WNG-FS85CJQ6-S
.
gaow@nenu.edu.cn
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ObjectType-Article-2
ObjectType-Feature-1
content type line 23
ISSN:0039-0402
1467-9574
DOI:10.1111/j.1467-9574.2011.00513.x