Abstract: 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.

Key words: interval type2 fuzzy sets, membership function (MF), data preprocessing, footprint of uncertainty (FOU), filled in