摘要
在数据挖掘领域,模糊C均值聚类法(FCM)在处理小量低维的数据挖掘时是有效的,但是面向数据库的数据挖掘经常要处理大量、高维的数据.在这种情况下,FCM算法在时间性能上难以令人满意.本文基于采样技术对FCM算法进行改进,以提高算法的时间性能,并利用遗传算法对聚类结果进行优化以保证聚类的质量,给出了一种新的基于遗传优化的采样模糊C均值聚类算法SFGO(SamplingFCMwithGeneticOptimization).仿真实验证明SFGO算法在大规模数据库的聚类挖掘中,在时间性能和聚类质量上都能获得较满意的结果.
In data mining field, FCM algorithm is an efficient method in the process of small scale low dimensional database, but the time performance of FCM algorithm can not be satisfied for the large scale high dimensional database. In this paper, a new sampling FCM algorithm with genetic optimization (SFGO) is presented based on the sampling technique and genetic algorithm. The sampling technique and genetic algorithm are used in SFGO algorithm to improve the time performance and the quality of clustering. The simulation experiment shows that the SFGO algorithm is an effective method in the data mining of large scale database.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2004年第5期121-125,共5页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(70273044)