气候变化对制造业产值的影响研究——基于非线性PLSR模型  

A Study on the Effects of Climate Change on the Manufacturing Output Based on Nonlinear PLSR Model

作　　者：陈纪波[1] 胡慧[1] 王桂芝[1] 

机构地区：[1]南京信息工程大学,南京210044

出　　处：《阅江学刊》2017年第3期72-80,共9页Yuejiang Academic Journal

基　　金：国家自然科学基金项目"基于‘互联网+’的中国制造业转型升级研究"(71673145);教育部哲学社会科学发展报告项目"中国制造业发展研究报告"(13JBG004)

摘　　要：为了研究气候因素对南京市制造业产值的影响情况,在HP滤波分离气象产值的基础上,考虑到气候因子间的多重共线性、气候因子与制造业产值之间复杂的非线性关系,尝试构造通过高斯核函数变换及三次B样条变换的气候因子对气象产值影响的非线性PLSR模型,并将其与常用的超越对数生产函数及普通PLSR模型进行比较。结果表明:该模型在进行气象产值模拟预测时误差相对较小,仅为347%,且高斯核函数变换的非线性PLSR模型更适用于南京市制造业气象产值的模拟预测。In order to study the impact of climate on the Nanjing manufacturing output,this article attempts to construct a nonlinear PLSR model, which is converted by Gauss KernelFunction and Cubic B-Spline Function, to show the impact of climatic factors on themeteorological output and then compare it to the Translog Production Function and generalPLSR model on the basis of separating the manufacturing meteorological output by HP filter.This is underpinned by the consideration that the multicollinearity among climatic factors andthe complex nonlinear relationship between climate factors and manufacturing output. Theresults show that, in terms of simulating predication of the meteorological output of themanufacturing industry, the Kernel-PLSR model and Spline-PLSR model in this article haverelatively smaller errors which only account for 3 47%. Furthermore, Kernel-PLSR model ismore applicable to predict the meteorological output of Nanjing manufacturing industry.

关 键 词：气象产值 高斯核函数 三次B样条函数 超越对数生产函数 PLSR模型 

分 类 号：F427[经济管理—产业经济] P467[天文地球—大气科学及气象学]