Determining whether a Belief Net is Consistent with Auxiliary Information
- Russ Greiner, Dept of Computing Science; PI of AICML
- Christian Darken
A belief net (BN), whether produced by interviewing domain expert(s) or learned from a body of data, can be incorrect. This report presents two tools for evaluating a given BN, by determining whether that network is statistically consistent with a body of accurate auxiliary information, IA. It also suggests how this information can be used to improve the BN.
Citation
R. Greiner, C. Darken. "Determining whether a Belief Net is Consistent with Auxiliary Information". Conditional Independence Structures and Graphical Models, Toronto, September 1999.| Keywords: | belief net |
| Category: | In Workshop |
BibTeX
@misc{Greiner+Darken:99,
author = {Russ Greiner and Christian Darken},
title = {Determining whether a Belief Net is Consistent with Auxiliary
Information},
booktitle = {Conditional Independence Structures and Graphical Models,
Toronto},
year = 1999,
}Last Updated: July 07, 2007Submitted by Russ Greiner
