主成份分析在城市生态经济动态评价中的应用

城市生态经济的动态发展过程,可以看作多种生态关系要素相互作用的集合沿时序轴波动的轨迹.在这种多变量分析中,完全采用人工筛选变量的方法难以避免确定权重等过程所带来的主观干扰.PCA(主成份分析)采用一种特殊的转换降维方法,根据原始数据自身的特点综合成几个主导因子,并最大程度反应原始数据的信息量.本文根据天津市四十年18项生态经济指标,采用相关矩阵和PCA双重筛选过程,得到载荷量大于90%的前三个主分量,将其作为天津市生态经济动态评价的基础.结果表明:天津市解放后四十年每十年一次的经济周期上迭加了自然社会等非经济因素的影响,表现出城市生态经济动态发展的三大阶段,这种发展主要受到宏观经济管理模式及其相应的投资机制的制约.

PRIMARY CONPONENT ANALYSIS APPLIED IN DYNAMIC EVALATE OF URBAN ECO-ECONOMY

Dynamic process of urban eco-economic development could be abstracted as a set of eco-variables moving along the time axis. It is very difficult to avoid subjective interferes in the multi- variable analysis by using the method of artificial screening. Primary Conponent Analysis ( PCA ) applied a special method of step-down variables is designed to find out the primary components of observation variables which prossess the dimension coordinate adopting the rotation transformation for coordinate axis, so that the sum of squares of error caused by descending the simension of the variable sec will be least.Three primary components have been chosen by using correlation coefficient matrix and PCA method according to 18 eco-economic variables from 1949 to 1988 in Tianjin.