Model-Based Knowledge Acquisition and Its Paradigm
基于模型的知识获取及其一个范式

Model-based knowledge acquisition is discussed and a paradigm is presented in this paper. Automatic acquisition of production rules of diagnostic expert systems is taken as examples. According to this paradigm,deep domain knowledge is represented by multivariate belief functions, while causes are obtained via abduction with uncertainties given by belief functions. This paradigm deserves robustness and nonmonotonicity.