Toward generating causal explanation: qualitative simulation with association mechanism to quantitative information

摘要

One of the central issues of qualitative reasoning is generating causal explanation in response to a user's query. The construction of an adequate model is crucial to the generation of a valid explanation in the case of large scale systems. This paper describes a new approach to the model construction based on using a causal ordering algorithm and qualitative conversion of numerical data. Our approach involves two key ideas: (1) a qualitative model is constructed from a unique quantitative model and (2) simulated qualitative behavior aspects are compared with numerical data. This approach is illustrated by using it to generate causal explanation related to the operation of a nuclear power plant.