| | |
Summary: Explanations in Bayesian Networks using
Provenance through Case-based Reasoning
Anders Kofod-Petersen, Helge Langseth, and Agnar Aamodt
Department of Computer and Information Science,
Norwegian University of Science and Technology,
7491 Trondheim, Norway
{anderpe|helgel|agnar}@idi.ntnu.no
Abstract. Bayesian Networks are useful for solving a wide range of
problems in many domains. Yet, they are exposed to one important
challenge when structural and parametrical changes occur. As Bayesian
networks lack memory regarding changes over time, there is currently
no good way of maintaining a history of changes and their provenance.
Thus, any variance in the network's problem solving behaviour will not
be explainable to a user. Within the context of systems that integrate
case-based reasoning and Bayesian networks, we suggest to add a case-
based reasoning functionality that will retain changes and their prove-
nance, as well as approaches to explain any unexpected problem solving
behaviour.
1 Introduction
Explanations have been identified as one of the most important properties of
|
