A study on generating fuzzy classification rules using histograms

We examine the performance of four approaches to the fuzzy rule generation for pattern classification problems. Two approaches generate a single fuzzy if-then rule for each class by specifying the membership function of each antecedent fuzzy set using the information about attribute values of training patterns. The other two approaches are based on fuzzy grids with homogeneous fuzzy partitions of each attribute. Since these four approaches are very simple and involve no time-consuming procedures, they can be easily implemented and applied to real-world pattern classification problems. The performance of each approach for test patterns (i.e., the generalization of ability of each approach) is evaluated by cross-validation techniques on commonly used data sets. Simulation results are compared with the performance of various classification methods reported in the literature.