【摘要】 以中国工业经济与环境的协调关系为例,利用相空间重构技术将一维时间序列扩展到多维,从而作为经遗传算法优化的神经网络的输入输出,由此对2005-2010年的中国工业与环境加以预测,使得预测精度大大提高。在此基础上,运用数据包络分析法对1986-2010年的工业经济与环境协调关系的有效性加以评价,指出二者之间大致呈“S”型发展趋势,表明中国工业环境污染状况正逐步得到改善。模型评价结果与实际发展基本相符,这为研究经济与环境协调发展提供了一个新的方向。更多还原

【Abstract】 Takingthe coordinated development between industry and environment in China as an example,this research expands one-dimension data to multi-dimension by using the method of phase space reconstruction that can be used as input and output of the genetic neural network.Based on these,the industry and environment have been forecasted from 2005 to 2010,and the accuracy that compared with other methodshas achieved significantly.Followed above,the relative validity of coordinated relationship between industry and environment has been evaluated from 1986 to 2010 by the method of data envelopment analysis.The results show that the coordinated relationship presents a "S" tendency between industry and environment which suggeststhat Chinese industrial pollution has been gradually improving.These conditions have been confirmed by the treatment rate of industrial pollutants and other researches.But this coordinated relationship is only relatively,it can be improved with development of industry.Therefore,from now on,it is important to perfect industrial structure,speed up technology innovation,strengthen management of heavy pollution enterprises,and reasonably adjust the spatial layout.Example indicates that the combination forecasting model integrated organically phase space reconstruction,genetic algorithm,artificial neural network and data envelopment analysis not only enhanced the forecast reliability,but also could be able to make further analysis,which will get researches more accurately,reasonable and objective.This combination forecasting model provides a new approachto study coordinated development.更多还原