Fast algorithm for iterative proportional fitting in log-linear models

This paper review the iterative proportional fitting algorithm for log-linear models and compares several implementations. Among these is a new proposal which is based on the application of sparse matrix techniques to the model matrix, and which exploits the special structure of hierarchical log-lin...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computational statistics & data analysis 1986-03, Vol.3 (4), p.251-264
Hauptverfasser: Denteneer, D, Verbeek, A
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper review the iterative proportional fitting algorithm for log-linear models and compares several implementations. Among these is a new proposal which is based on the application of sparse matrix techniques to the model matrix, and which exploits the special structure of hierarchical log-linear models. It is very efficient, both in storage and in execution time.
ISSN:0167-9473