中国产业系统非期望产出率熵变层级结构关系研究——基于PLS回归的层级因素分解技术  

Hierarchy Relationship of Undesirable Output Rate Entropy Change in China Industry System: Based on Hierarchical Factor Decomposition Technique with PLS

作　　者：李恒川[1] 路正南[1] 

机构地区：[1]江苏大学管理学院,江苏镇江212013

出　　处：《中国人口·资源与环境》2015年第8期116-124,共9页China Population,Resources and Environment

基　　金：国家自然科学基金项目"基于非期望产出率熵变的产业系统减排机理及碳减排差别责任研究"(编号:71173094)

摘　　要：产业系统具备开放系统、远离平衡态以存在涨落的耗散结构特征,以此为基础可以构建我国产业系统的熵变理论,从产业系统熵变路径的角度出发来探讨产业系统的减排途径。产业系统熵主要用于说明产业系统状态,产业系统熵值不能体现产业系统熵变的情况,所以为了探寻产业系统熵变路径,需要运用层级因素分解技术对非期望产出率熵变进行两层分解。使用基于kaya等式的LMDI分解法对非期望产出率熵变进行第一层分解,将其分解为产业结构熵变、非期望产出强度熵变和期望产出率熵变,再参考kaya等式的形式,引入生产函数作为因素分析的要素依据,对第一层分解出的因素进行第二层分解,梳理从最终产品需求以及中间技术变化出发,经由产出结构变动、期望产出率变动、非期望产出强度变动最终影响产业系统非期望产出率熵变的网络层级结构,勾勒出产业系统非期望产出率熵变的路径图,利用路径分析模型分析各影响因素之间的网络层级结构关系。基于2000-2012年我国产业部门的数据,本文对产业系统非期望产出率熵变进行了实证分析,在此过程中,利用PLS回归对我国产业部门的生产函数进行了拟合,解决了生产函数拟合过程中存在的多重共线性问题,为非期望产出率的熵值计算做好了数据准备。通过计算我国产业系统非期望产出率及各分解因素的熵值,结合路径分析模型的结果,可以得到结论:我国非期望产出率熵变的涨落与我国经济发展的周期吻合,验证了熵变理论的正确性和实用性;从路径分析模型的总效应角度分析,提高生产过程中各生产要素的使用效率对我国产业系统的熵减贡献较大;路径分析模型的直接和间接效应都显示产业结构对非期望产出率变动影响最大,产业结构转型升级是实现产业系统减排的重要途径。本文的研究结论为实现我国产业系The industrial system is an open system far away from equilibrium and fluctuates sometimes. It means that the industrial system has the characteristics of dissipative structure, which can be the basis of the construction of China＇ s industrial system entropy theory. We can explore the industrial system reduction approaches with the industrial system entropy change path. The industrial system entropy mainly explains the state of industrial system and does not presents the change of the entropy. To investigate the change path of the industrial system entropy, we require a hierarchical factor decomposition technique by which we decompose the industrial system entropy into two layers. This paper uses the LMDI decomposition based on Kaya equation to carry on the first layer decomposition, and decomposes the undesirable output rate entropy into the industrial structure entropy, the undesirable output intensity entropy and the expected output rate entropy. In the second layer decomposition, the paper introduces the production function into the decomposition method just like Kaya equation. After the hierarchical factor decomposition, We can obtain a hierarchical network structure with using the method of path analysis to discuss the entropy change, which begins from the middle technological changes and the final product demand, then affects the change of the output structure, the rate of expected output, and the intensity of undesirable output intensity, and finally affects the entropy change of industrial system. We carry on an empirical analysis based on the relevant data of industrial system undesirable output rate of the industrial sector in China in 2000 - ：2012. Considering the multicollinearity problems of the production function, we use a PLS regression model to fit it. The result prepares data for calculating our industrial system entropy. With analysis of the industrial system entropy by using the path model, we draw following conclusions： the undesirable output rate entropy corresponds to the economy cyc

关 键 词：非期望产出率 熵变 层级分解 路径分析 PLS回归 

分 类 号：F206[经济管理—国民经济]