Fuzzy Type-1 Triangular Membership Function Approximation Using Fuzzy C-Means

Abstract

Fuzzy logic is a way of many-valued computing logic that deals with the truth values of the variables between 0 and 1, unlike the conventional Boolean logic. Membership functions are used to depict the fuzzy values of given variable. Though membership functions are determined through expert’s opinion, however, the one estimated through heuristic algorithms is the preferable methods. Membership functions determined through statistical and knowledge engineering methods are usually application dependent and cannot be applied on different datasets. This research focuses on generating the parametric values of the triangular membership function using a novel method. Initially, the Fuzzy C-means algorithm is utilized to generate the parameters values of the Gaussian membership function. Using a set of equations, these values then estimate the parameters of the triangular membership function. The proposed method is applied to the quality of web services data. From the results it is verified that the new approach of generating triangular membership functions can be adopted.