【摘要】 提出了一种基于遗传进化策略的聚类分析新方法，将聚类问题转换为实数空间的优化问题．用遗传策略在连续的实数空间寻找最优类中心，既可避免求解组合优化问题，减少计算量，又可避免在优化过程中陷入局部最优．此遗传聚类方法用于对4组实际数据进行硬分类和模糊分类，结果令人满意．更多还原

【Abstract】 Genetic evolution strategy (GEs) was used in chemical clustering analysis. The clustering objective was formulated as a function of real-valued parameters. Combinatorial optimization usually encountered in clustering analysis, which is commonly time-consuming in computation, can be avoided, and therefore, the amount of computation can be significantly reduced. Global optimum might be achieved with the help of genetic algorithm. Results obtalned from four real chemical data sets substantiate the practicability of the method proposed in the paper.更多还原