Explanation based on contexts

Abstract

The authors present a hypertext-based explanation mechanism built on an existing diagnostic expert system. The system is a decision support tool for experts analyzing the condition of rotating machinery. Explanation for the system is based on context. It guides interaction and search for suitable explanations. Questions and answers are modeled separately, allowing heuristic answering. System knowledge is modeled as explicable objects to make explanations more modular. Using context to maintain a dialogue is discussed with an implementational view, and explanation techniques of an actual real-world system are presented.<>