论文部分内容阅读
产生式规则是智能系统中大量采用的一种知识表示模式。规则常带有不确定性,规则前提中的变量值在很多情况下是一个模糊集,而当前事实中的变量值可能是模糊集、实数域上的区间值或者清晰数。为了统一处理这些情形,提出了基于证据理论,通过对两种贴近度进行加权合成计算规则匹配度的方法。该方法可根据变量值的分布特点自适应地调整每种贴近度的权重,从而使规则的匹配结果更为合理。
Production rule is a kind of knowledge representation mode widely used in intelligent system. The rules often have uncertainties. The value of a variable in the rule premise is a fuzzy set in many cases. The current variable value in fact may be a fuzzy set, an interval value in a real number field, or a sharp number. In order to deal with these cases uniformly, a method based on the evidence theory is proposed to calculate the degree of rule matching by weighted synthesis of two kinds of closeness. This method can adaptively adjust the weight of each close degree according to the distribution of the variable values so as to make the rule matching result more reasonable.