Genetic lateral tuning of membership functions as post-processing for hybrid fuzzy genetics-based machine learning

摘要: Genetic fuzzy systems (GFS) have been actively studied in the field of classifier design. GFS can generate simple and accurate classifiers with a number if-then rules by evolutionary computation (EC). One general concern is how to discretize numerical attributes into partitions. Most use homogeneous partitions without considering class distribution each attribute. This simplest idea, but more could be obtained optimizing There are three approaches. pre-processing where inhomogeneous specified according before applying EC. Another approach simultaneously optimize both set The other post-processing used optimized afterward. In this paper, we examine effect partition optimization applied GFS. computational experiments, first our hybrid genetics-based machine learning for standard data sets its parallel distributed implementation large sets. Then apply genetic lateral tuning as shift positions membership functions pattern distribution.