基于增强蜂群优化与K-means的文本聚类算法  

Enhanced bee colony optimal and K-means based document clustering algorithm

作　　者：柯钢[1] 

机构地区：[1]东莞职业技术学院计算机工程系,广东东莞523808

出　　处：《计算机应用研究》2016年第8期2298-2302,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61106019);东莞市社会科技发展项目(2013108101045)

摘　　要：针对文本数据维度较高、空间分布稀疏及其聚类效果不佳的问题,提出一种基于增强蜂群优化搜索与K-means的高效文本聚类算法。首先为蜂群算法引入公平操作与克隆操作来提高全局搜索的能力,公平操作提高了样本多样性,并增强了蜂群搜索能力;克隆操作则增强了各代之间的信息交流,提高了求解质量。最终引入K-means进行局部质心的提炼,提高聚类质量。基于文本数据集的实验结果证明,相较于其他聚类算法,本算法具有更高的聚类质量。Aiming at the problem that the document data has the characteristics such as high-dimensionality and sparseness, this paper proposed an enhanced bee colony optimal and K-means based document clustering algorithm. Firstly, it introduced the fairness and clone operation to bee colony to improve the global search power, enhanced the individuals diversity and the search power by fairness operation, enhanced the information communication between different iterations by clone operation, and improved the solution quality. At last,it improved the clustering quality by K-means which was good at local refining. Ex- periments results based on the documents show that the proposed algorithm has better clustering quality than the other cluste- ring algorithm.

关 键 词：蜂群算法 公平操作 克隆操作 多样性 局部提炼 文本聚类 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]