作者: Tzung-Pei Hong , Jyh-Bin Chen
DOI: 10.1016/S0165-0114(97)00187-5
关键词:
摘要: Fuzzy systems that automatically derive fuzzy if-then rules from numeric data have been developed. Most to predefine membership functions in order learn. Hong and Lee proposed a general learning method derives set of given training examples using decision table. All available attributes were included the table initial for each attribute built according predefined smallest unit. Although Lee's accurately final functions, are complex if there many or unit is small. We improve by first selecting relevant building appropriate functions. These then used Experimental results on Iris show effectively induces rules.