基于细菌群体趋药性优化的k-means算法  

K-means algorithm optimization based on bacterial colony chemotaxis

作　　者：于来行[1] 张敏[1] 葛斌[1] 

机构地区：[1]大连大学信息工程学院,辽宁大连116622

出　　处：《计算机工程与设计》2009年第7期1725-1727,1737,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(60573072)

摘　　要：细菌趋药性算法是一种新的仿生进化算法,针对细菌趋药性算法,介绍了其基本原理,讨论了一种改进的算法——细菌群体趋药性(BCC)算法。将细菌群体趋药性优化方法应用在k-means聚类分析中,以类内离散度和为目标函数,建立了BCC优化的k-means算法模型,利用BCC算法的全局搜索能力,很大程度上避免了k-means算法易陷入局部极小的缺陷,同时也降低了算法对初始值的敏感度。并给出了一些实验,结果令人满意。Bacterial chemotaxis optimization is a novel category of bionic algorithm for optimization problems. Firstly, the basic principle of bacterial chemotaxis optimization is introduced, and then a improved method is discussed that is bacterial colony chemotaxis algorithm. Bacterial colony chemotaxis optimization algorithm is used in k-means clustering analysis. The modeling of k-means algorithm is established taking the sum with-in cluster scatter as the objective. Using the BCC algorithm global search capability, avoids the local minimum problems of k-means algorithm. At the same time, algorithm is no longer a large degree dependent on the initialization values numerical. Exoerimental results are satisfied.

关 键 词：细菌趋药性算法 细菌群体趋药性 K-MEANS算法 聚类算法 类内离散度和 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]