摘要
为解决文档聚类问题,提出一种基于差分进化的聚类算法,通过把文档聚类问题建模为优化问题,对聚类准则函数进行优化,来寻找初始最优聚类中心.在此基础上,进一步提出两种差分进化算法与K均值结合的混合方法,来获得更好的聚类结果.实验表明,与经典K均值算法相比,新提出的两种混合方法能够获得较好的聚类质量.
This paper proposes a novel differential evolution clustering algorithm for solving Web document clustering.First,by modeling Web document clustering problem as an optimization problem,the clustering criterion function is optimized aiming at finding the promising initial centroids.Then,K-means and differential evolution clustering algorithm are hybridized in two ways to achieve better clustering performance.Compared with K-means algorithm,experimental results reveal that the two proposed hybrid approaches can acquire better and higher clustering quality.
出处
《聊城大学学报(自然科学版)》
2012年第1期93-97,共5页
Journal of Liaocheng University:Natural Science Edition
基金
山东省教育厅科技计划项目资助(J08LJ59)
关键词
WEB文档聚类
差分进化算法
K均值
优化问题
Web document clustering
differential evolution algorithm
K-means
optimization problem