电子学报 ›› 2016, Vol. 44 ›› Issue (2): 405-412.DOI: 10.3969/j.issn.0372-2112.2016.02.023

• 学术论文 • 上一篇    下一篇

基于量子谐振子模型的聚类中心选取算法

燕京京1,3, 王鹏2, 范家兵1,3, 黄焱1,3   

  1. 1. 中国科学院成都计算机应用研究所, 四川成都 610041;
    2. 成都信息工程学院并行计算实验室, 四川成都 610225;
    3. 中国科学院大学, 北京 100049
  • 收稿日期:2014-07-01 修回日期:2014-12-06 出版日期:2016-02-25 发布日期:2016-02-25
  • 通讯作者: 王鹏
  • 作者简介:燕京京 女,1989年生于河南安阳.硕士研究生,主要研究领域为数据挖掘.
  • 基金资助:

    国家自然科学基金(No.60702075);广东省科技厅高新技术产业化科技攻关项目(No.2011B010200007);四川省青年科学基金(No.09ZQ026-068);成都市科技局创新发展战略研究项目(No.11RXYB016ZF)

Clustering Center Selecting Algorithm Based on Quantum Harmonic Oscillator Model

YAN Jing-jing1,3, WANG Peng2, FAN Jia-bing1,3, HUANG Yan1,3   

  1. 1. Chengdu Institute of Computer Application, Chinese Academy of Sciences, Chengdu, Sichuan 610041, China;
    2. Parallel Computing Laboratory, Chengdu University of Information Technology, Chengdu, Sichuan 610225, China;
    3. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2014-07-01 Revised:2014-12-06 Online:2016-02-25 Published:2016-02-25

摘要:

提出了一种基于量子谐振子模型的聚类中心选取算法.该算法以量子谐振子波函数从高能态到基态过程中的概率变化过程为理论模型来描述聚类问题中数据对象向聚类中心点的聚集行为,能够快速查找到最优的聚类个数及较好的聚类中心点所在的网格;数据读入网格结构之后,算法的处理时间与数据集规模无关.实验结果表明:CCSA-QHOM算法较适合于处理每个子类局部区域的网格密度分布呈单峰特性的数据集的聚类中心选择问题.

关键词: 聚类中心, 量子谐振子, 聚类个数, 网格, 单峰特性

Abstract:

This article puts forward a clustering center selecting algorithm based on quantum harmonic oscillator model (CCSA-QHOM).The algorithm describes the way of data objects finding center of the cluster in clustering problem by taking the change of wave function's probability in the process of high energy level to a lower energy level for theoretical model.It can quickly find the optimal number of clusters and cluster center, computing time has nothing to do with the size of the data set after the dataset being got in grid space.Experiments show that CCSA-QHOM is more suitable for processing the clustering center selection question of dataset in which grid density distribution of each subclass haves a single peak characteristic.

Key words: cluster center, quantum harmonic oscillator, number of clusters, grid, peak characteristic

中图分类号: