基于粗糙集的不确定知识表示方法

Uncertainty Knowledge Representation Method Based on Rough Sets

1 引言知识表示是人工智能(AI)研究中最关键的分支之一。传统的知识表示模型(如AQ11,ID3等),对知识描述是确定的、清晰的,即:被描述的对象具有或不具有某种属性是明确的。然而,在现实世界中,人们常常面对的是在领域信息不完整、不确定、不精确的前提下,完成对事物的认识、分析、推理、判断、预测和决策。这种智能行为往往要求人们对未知的信息进行估计、推测;对不完整数据进行分析、处理;

A property set model for knowledge representation is proposed by discussing knowledge str-cuture and representation mechanism, and defining property set and atomic concept based on rough set theory. Uncertainty knowledge can be represented by upper and lower approximation, and certainty knowledge can be also represented exactly by using the model. The model presents a new idea to uncertainty knowledge representation.