文档介绍:基于量子遗传谱聚算法的聚类
蒋勇1,谭怀亮2
(,湖南,株洲,412004;
2. 湖南大学计算机与通信学院,湖南,长沙,410082)
(hunanlaojiang@)
摘要:主要核方法研究XML聚类,提出了一种改进的XML文档核聚类方法。该方法先对XML文档约简,以频繁标签序列建立向量空间核的核矩阵,用高斯核函数求解初始聚类和聚类中心,然后用初始聚类中心构造量子遗传算法的初始种群,通过量子遗传算法与核聚算法相结合求得全局最优解的聚类。为了验证本文提出的算法,实验结果显示,使用该算法的聚类比改进的核聚算法、K—means等单一方法具有良好的收敛性、稳定性和更高的全局最优。
关键词:XML文档;高斯核函数;核聚类算法;量子遗传算法;XML聚类
中图法分类号: 文献标识码:A
Clustering Based on Quantum ic Spectral Clustering Algorithm
JIANG Yong1,TAN Hui-liang2
( of Information and College of Hunan Chemical ,Zhuzhou Hunan 412004,China
puter munication,Hunan University,Changsha Hunan ,410082,China)
Abstract:This paper mainly targets on XML Clustering with kernel methods for pattern analysis and the quantum ic algorithm, A new method based on the quantum ic algorithm and clustering algorithm was derived. To the XML documents eliminated, the vector space kernel’s kernel matrix were generated with frequent-tag sequence , first solves the initial clustering and clustering center with the Gaussian kernel functions , then the quantum ic algorithm's initial populations were constructed by the initial clustering center structure, clustering of the globally optimal solutions were obtained through it and kernel clustering algorithm. In order to confirm the algorithm which this article proposed, the experimental result showed that it is more superior to the improvement