NEGATIVE AND POSITIVE EXPLANATIONS IN EXPERT SYSTEMS

Abstract
This paper presents an explanation module implemented in the Computer Science Department of the University of Orsay. It was designed to be used after a session with an expert system and to answer two types of questions: why does some fact have a particular value (deduced by the expert system in the session) and why does some fact not have a particular value (after the expert system has given a value different from the one the user expected) ? The answers to these questions are referred to as positive and negative explanations, respectively.

This publication has 1 reference indexed in Scilit: