地理学报 ›› 2009, Vol. 64 ›› Issue (12): 1439-1447.doi: 10.11821/xb200912004

• 论文 • 上一篇    下一篇

广义多维云模型及在空间聚类中的应用

邓羽1,2,  刘盛和1,  张文婷3,  王丽2,4,5,  王江浩1,2   

  1. 1. 中国科学院地理科学与资源研究所,北京 100101;2. 中国科学院研究生院,北京 100049;
    3. 武汉大学资源与环境科学学院,武汉 430079; 4. 中国科学院科技政策与管理科学研究所,北京 100190;
    5. 中国科学院自然与社会交叉科学研究中心,北京 100190
  • 收稿日期:2009-01-09 修回日期:2009-09-05 出版日期:2009-12-25 发布日期:2009-12-25
  • 通讯作者: 刘盛和 (1967-), 男, 湖南衡阳人, 博士, 研究员, 研究方向为城市发展与土地利用。E-mail: liush@igsnrr.ac.cn
  • 作者简介:邓羽 (1985-), 男, 湖北恩施人, 硕士研究生, 从事城市发展与土地利用研究。E-mail: rain00788@163.com
  • 基金资助:

    国家自然科学基金项目 (40971102; 40871179); 中科院研究生科技创新与社会实践资助专项

Spatial Clustering Method Based on General Multidimensional Cloud Model

DENG Yu1,2,  LIU Shenghe1,  ZHANG Wenting3,  WANG Li2,4,5,  WANG Jianghao1,2   

  1. 1. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China;
    2. Graduate University of Chinese Academy of Sciences, Beijing 100049, China;
    3. School of Resources and Environment Science, Wuhan University, Wuhan 430079, China;
    4. Institute of Policy and Management, CAS, Beijing 100190, China;
    5. Center for Interdisciplinary Studies of Natural and Social Sciences, CAS, Beijing 100190, China
  • Received:2009-01-09 Revised:2009-09-05 Online:2009-12-25 Published:2009-12-25
  • Supported by:

    National Natural Science Foundation of China, No.40971102; No.40871179; The CAS Special Grant for Postgraduate Research, Innovation and Practice

摘要:

传统的空间聚类方法难于脱离“硬划分”的束缚,且不能有效描述空间对象的复杂特征。一维云模型无法准确反映现实世界的多属性特征,简单的数据融合丢失了空间对象的必要信息。标准二维云模型克服了一维云的不足,但是在模拟复杂地理现象的非齐性和非对称性方面显得捉襟见肘。基于以上考虑,提出了广义多维云模型,以分段特性来体现空间对象的综合特征,并推导出模型的数学表达式。在实证研究的基础上,从空间聚类的隶属程度空间分布特征、与模糊C均值的对比研究及与住宅地价的耦合分析三个视角,详实解读了聚类结果。分析发现,广义多维云模型更能体现空间对象的综合特征,空间聚类结果能够反映出空间分布的潜在信息,更为准确的实现了复杂情况下的空间划分。该模型在刻画地理现象中更为合理,但由于地理实体的空间作用极其复杂,建立模型是一项既具体又充满挑战的任务。

关键词: 多维云, 空间聚类, 数据挖掘, 隶属度

Abstract:

Traditional spatial clustering methods can not avoid the disadvantage of "hardware division", and can not describe the physical characteristics of spatial entity effectively. One-dimensional cloud model can not accurately reflect multi-attribute characteristics of the real-world. Besides, essential information of spatial objects might be lost during procedure of simple fusion. Standard two-dimensional cloud model overcomes some shortcomings of one-dimensional cloud, but it still can not meet the needs of simulating the non-homogeneous and non-symmetry characteristics of complex geographical phenomena. In view of the above, this paper sets forth a general multi-dimensional cloud model, which describes the characteristics of spatial objects more reasonably. Based on the empirical research, a detailed interpretation of clustering results is made from the spatial distribution of membership degree of clustering, the comparative study of Fuzzy C-means and a coupled analysis of residential land prices. It is found that general multi-dimensional cloud model can reflect the integrated characteristics of spatial objects better, reveal the spatial distribution of potential information, and realize spatial division more accurately in complex circumstances. However, due to the complexity of spatial interactions among geographical entities, the construction of cloud model is a specific and challenging task.

Key words: multi-dimensional cloud, spatial clustering, data mining, membership degree