摘要
不精确推理是根据应用系统中的不确定信息(可能是概率的、模糊的、不完备的等),按一定的搜索策略(如深度优先或广度优先),采用一定的推理技术(如正向、反向或双向)和不确定值传播算法,得出应用系统中近乎合理的结论。一个不精确推理模型主要包括知识的不确定性描述方式。
This paper developes an interval-valued-based approximate reasoning model, which includes different kinds of approximate reasoning methods, such as those based on probability, fuzzy set or evidence theory etc, and provides a single framework for dealing with different types of uncertainty informatibn, which may exist in application systems. This model bases on the idea of evidence-theory. It uses an interval value to describe the degrees of certainty, uncertainty and unknown of the uncertainty information. Finally, some examples of this model using in expert system and dealing with default knowledge are given.
出处
《计算机科学》
CSCD
北大核心
1989年第4期21-24,8,共5页
Computer Science
