A Method To Determine The Interval Type-2 Membership Function Via Axiomatic Fuzzy Sets And Its Applications

Membership functions are quantitative descriptions of fuzzy concepts.It is a fun-damental for solving practical problems to determine membership functions.Essentially,the membership functions are also in line with external rules that the people know the business and holds by the objective laws.However due to the cognitive differences of each person to fuzzy concepts,membership functions are inevitably empiricism and sub-jectivity.In order to fully reflect the objective laws and improve the ability of membership functions to deal with natural languages and uncertainty issues,the following researches are explored in this paper.1.Interval type-2 membership functions based on axiomatic fuzzy sets(AFSIT2 MF)are constructed.AFSIT2 MFs are determined by the interval value of the variance and the fixed mean value of the observed dates.This paper takes a fuzzy concept in the Iris data set as an example to illustrate the process of determining AFSIT2 MF.2.To verify the validity and rationality of AFSIT2 MF,a clustering algorithm based on AFSIT2 MF is constructed.Experimenting on four data sets,the experimental results not only show the effectiveness and rationality of AFSIT2 MF,but also illustrate the ability of AFSIT2 MF to explain?3.To verify the ability of AFSIT2 MF to deal with uncertain problems,A clas-sification algorithm based on AFSIT2 MF is proposed.12 data sets that from UCI are employed compare to with the classification algorithm based on AFS and classical clas-sification algorithms.The experimental results that the classification algorithm based on AFSIT2 MF has better classification accuracy and semantic interpretability,illustrate the ability of AFSIT2 MF to deal with uncertain problems.