An Approach to Automatic Generation of Fuzzy Membership Functions Using Popularity Metrics


Creating membership functions for fuzzy system can be a difficult task for non-expert developers. This is even more difficult when the information available about the specific domain is limited. In our case, we wanted to create membership functions that model the different characteristics of mobile devices. Due to the lack of public data about the mobile phones sales it is difficult to estimate the market share of each device. To tackle this problem we have developed a mechanism that uses popularity metrics to estimate the market share and generate the membership functions. In this paper we describe the used algorithm and discuss the obtained results.