摘要： 提出一种基于跨世代异物种重组大变异遗传模型的数量属性模糊划分方法，采用实数编码和PNX交叉算子，通过设定阈值重新初始化算子，以模糊1-频繁项的支持度、三角形隶属度函数对数量属性取值范围的覆盖率以及隶属度函数间的重叠度为优化目标，通过遗传进化自动确定数量属性的模糊划分区间。实验结果表明该方法缩短了进化所需时间，所得最优个体的适应度值较高。

关键词: 跨世代异物种重组大变异遗传模型, 模糊划分, 模糊关联规则, 模糊集合, 实数编码

Abstract: This paper presents a quantitative attribute fuzzy partition method based on Cross generation Heterogeneous recombination Catacly- smic mutation(CHC) genetic model. It adopts real number coding and PNX crossover operator, initializes operator renewedly by setting threshold value. Fuzzy 1-frequent item’s support degree, the coverage ratio of triangular membership functions for quantitative attribute and the overlap degree among membership functions are regarded as optimization object. It confirms automatically quantitative attribute’s fuzzy partition interval by genetic evolution. Experimental results show that the method reduces evolution time and the obtained optimal unit has better fitness degree value.

Key words: Cross generation Heterogeneous recombination Cataclysmic mutation(CHC) genetic model, fuzzy partition, fuzzy association rule, fuzzy set, real number coding

中图分类号: 

• TP391