PLS路径模型在建立综合评价指数中的应用

提出将PLS(partialleastsquare)路径模型和复数据表分析相结合,建立多指标系统评估指数的方法.该方法一方面可以方便地研究每一个隐变量与其显变量集合之间的关系,并可以分析各隐变量与综合评估指数之间的联系;另一方面,还可以充分提取原始变量中有效信息,得到一个既能够综合各隐变量、又能很好地代表系统中所有指标变量的综合指数.论文以中国城市发展水平评估为例,采用了反映经济水平、生活水平和城市化水平的三个变量组,将PLS路径模型用于建立城市综合评估指数,对不同城市进行综合评价排名,得到的结果与一般的经济地理知识是十分吻合的.

The Application Research of Partial Least Square Path Modeling on Establishing Synthesis Evaluation Index

In this paper, a method for establishing multivariable evaluation index is proposed by combining Partial Least Square Path Modeling with Multiple Tables Analysis. The relation between every latent variable and corresponding observable variables is investigated and the relation between every latent variable and multivariable evaluation index is analyzed with the proposed method. On the other hand, the important information of observable variables is extracted and the multivariable evaluation index which represents the latent variables and observable variables appropriately is found with the method. As a case research, three variable sets which reflect economic development level, life level and urbanization level are adopted and partial least square path modeling is applied for establishing city synthesis evaluation index on city development level of China. Through the evaluation of different city, it is consistent between the result and economic geography knowledge.