一种区间二型模糊隶属度函数的构造新方法

针对现有区间二型模糊隶属度函数的构造方法中，存在受试者需为模糊专家，易出现不确定迹(FOU)填满现象(下隶属度函数近似为0)，或需预先确定不确定迹的形状的问题，设计受试者不为模糊专家也可正确回答的调查问题，提出一种构造区间二型模糊集隶属度函数的新方法.该法不需预先确定FOU的形状，受试者不必为模糊专家.该法的基本思想是从一组领域专家中收集类属词的上下隶属度区间数据，对所获数据进行数据预处理，最后获得类属词的隶属度函数和FOU.实验结果表明，所提出的新方法可有效构造区间二型模糊隶属度函数，避免出现填满现象.

A New Construction Method of Membership Function for Interval Type-2 Fuzzy Sets

There are several problems in the construction method of membership function for interval type-2 fuzzy sets, such as it requires that subjects are knowledgeable about fuzzy sets, footprint of uncertainty (FOU) is filled in or almost filled in (lower membership function is approximately zero), and the shape of an FOU must be chosen ahead of time. In order to solve these problems, a novel construction method of membership function for interval type-2 fuzzy sets was proposed based on a question design for subjects who are not limited to experts of fuzzy sets. The proposed method did not need to choose the shape of a FOU. The basic idea of the method could be concluded as follows. First, the interval data for upper membership degree and lower membership degree of a word was collected from domain experts. Then, the interval data would be preprocessed. And at last, the membership function and footprint of uncertainty for the word would be obtained. Experimental results showed that the proposed method can establish membership function for interval type-2 fuzzy sets effectively and can avoid the FOU to be filled in or almost filled in.