A new scheme is proposed to estimate the membership values for fuzzy sets. The scheme takes input from empirical/experimental data which reflects the expert's knowledge on the relative degree of belonging of the members. It then searches for the best fit
