Abstract: Index weight is important in retrieval model of Case-Based Reasoning(CBR) system. Index weight was worked out by the method of two-level fuzzy synthesis evaluation based on several experts. Deviation degree and consistency of each expert were analyzed by relation degree. It is proved to be effective and feasible by experiments.

Key words: Case-Based Reasoning(CBR), two-level fuzzy synthesis evaluation, deviation degree, relation degree

摘要： 指标权重的确定在基于实例推理(CBR)系统的检索模型中起着重要的作用。采用基于多位专家的二级模糊综合评判方法求得各个指标的总的综合权重,对指标权重进行了讨论,并引入关联度的概念,讨论了各专家的偏离度及一致性。实例证明有效、可行。

关键词: 基于实例推理(CBR), 二级模糊综合评判, 偏离度, 关联度