地理研究 ›› 2001, Vol. 20 ›› Issue (6): 761-767.doi: 10.11821/yj2001060014

• 论文 • 上一篇    

典型相关分析在土地利用结构研究中的应用——以环渤海地区为例

张明, 朱会义, 何书金   

  1. 中国科学院地理科学与资源研究所, 北京 100101
  • 收稿日期:2001-09-08 修回日期:2001-10-26 出版日期:2001-12-15 发布日期:2001-12-15
  • 作者简介:张明(1970-),男,河南安阳人,博士。主要从事土地利用与土地覆被变化研究。
  • 基金资助:

    国家自然科学基金项目(49831020);中科院知识创新工程项目(KZCX2-310-01-06)

An application of canonical correlation analysis to research on land-use patterns ——Taking Bohai Rim as an example

ZHANG Ming, ZHU Hui-yi, HE Shu-jin   

  1. Institute of Geographic Sciences and Natural Resources Research, Beijing 100101, China
  • Received:2001-09-08 Revised:2001-10-26 Online:2001-12-15 Published:2001-12-15

摘要:

典型相关分析是一种揭示两组多元随机变量之间相关关系的统计模型方法。本文通过介绍这一多元统计方法的内涵、特点和思路,将其引入有关地学问题的分析———判别土地利用类型分布与其影响因子之间的相关关系。以环渤海地区为例,通过数据准备、操作过程和统计检验等几个方面,全面介绍了典型相关分析在实际研究中的应用

关键词: 土地利用, 典型相关分析, 环渤海地区

Abstract:

Land-use and land-cover change (LUCC) has become a key field for the research of global environmental change since the 1990s. More attention is paid to the identification and description of the driving forces of LUCC. For all kinds of limitations, it is difficult to incorporate different influencing factors into a whole model to carry out such an analysis. Canonical Correlation Analysis (CCA) is a kind of traditional statistical model to show the relationship between two multiple stochastic variables. It is widely applied in different research fields to describe the relationship between the set of criterion measures and the set of explanatory factors. This paper introduces the connotations, characteristics and procedures of this multivariate analysis. It is combined with one of the analyses on geographic problems, to identify the relationship between the land-use patterns and its influencing factors. Taking the Bohai Rim in China as an example, through the preparation of the data, operation procedures and statistical checks, the application of CCA in a practical study is given comprehensively. It is shown that CCA could be widely applicable to the geographic analysis, especially to the identification of the relationship between land-use structure and its influencing factors. Through the canonical loadings analysis, the influencing extension of different factors to the dependant variables or the contribution extension of explanatory variables to the explained variables could be described quantitatively. As CCA is carried out in combination with a series of data checks and precision tests, it makes the specific analysis more rational and scientific.

Key words: land-use change, canonical correlation analysis, Bohai Rim

  • F301.24