Relaxing the Local Independence Assumption for Quantitative Learning in Acyclic Directed Graphical Models through Hierarchical Partition Models

Authors:

Daniela Golinelli
Dept. of Statistics, Box 354322
University of Washington
Seattle, Washington, 98195
E-mail: golinell@stat.washington.edu
Phone: 206-543-4302
Fax: 206-685-7419

David Madigan
Dept. of Statistics, Box 354322
University of Washington
Seattle, Washington, 98195
E-mail: madigan@stat.washington.edu
Phone: 206-543-4537
Fax: 206-685-7419

Guido Consonni
Dip. di Economia e Metodi Quantitativi
Universita' di Pavia, Via S. Felice, 5
27100 Pavia, Italy
E-mail: gconsonni@eco.unipv.it
Phone: +39-0382-506-225
Fax: +39-0382-304226

Abstract:

The simplest method proposed by Spiegelhalter and Lauritzen (1990) to perform quantitative learning in ADG presents a potential weakness: the local independence assumption. We propose to alleviate this problem through the use of Hierarchical Partition Models. Our approach is compared with the previous one from an interpretative and predictive point of view.

Keywords:

Local independence assumption, hierarchical partition models, Bayesian networks

Availability:

PostScript