基于KPCA的生态效率评价与预测模型研究

发布时间：2018-05-01 14:09

本文选题：生态效率 + 评价与预测　；参考：《华北理工大学》2017年硕士论文

【摘要】：随着社会经济的发展,经济增长与资源环境的矛盾日益凸显,为了协调经济与资源环境的关系,开展生态效率的相关研究具有重大意义。首先利用非线性回归与信息熵针对核主成分分析法应用中核函数和核参数的选取进行了探讨,基于UCI和统计年鉴原数据集拟合信息熵与核函数参数的经验公式22?(10)(28)291.11274-410.21300612.9958-xx。其次在归纳综合了生态效率指标选择的目标及原则的基础上,针对目前国内外对生态效率的研究以及研究存在的问题,结合当前经济发展与资源环境的关系,对原有的生态效率评价、预测指标体系进行了改进。利用核主成分分析法、数据包络分析法,对经典的生态效率评价模型进行改进,建立了KPCA-DEA复合评价模型。接着,应用核主成分分析法、BP神经网络预测法等方法对经典预测模型进行改进,建立了KPCA-BP复合预测模型。实证分析结果显示,DEA模型评价的综合技术效率值有14个为1,PCA-DEA模型的综合技术效率值为1的有12个,而KPCA-DEA模型综合技术效率值有6个为1,从综合技术效率值结果表明,KPCA-DEA复合评价模型的效果优于PCADEA和DEA。说明生态效率复合评价中具有一定的优势和可行性。KPCA-BP复合预测模型实证分析显示,KPCA-BP的相对误差为1.56%,PCA-BP相对误差为17.15%,而BP模型预测的相对误差为20.75%,结果表明KPCA-BP复合预测模在预测中具有一定的优势。通过本研究丰富了生态效率的评价和预测方法,为生态效率的评价与预测提供了参考。
[Abstract]:With the development of social economy, the contradiction between economic growth and resource environment is becoming increasingly prominent. In order to coordinate the relationship between economy and resources and environment, it is of great significance to carry out the research on ecological efficiency. Firstly, using nonlinear regression and information entropy, the selection of kernel function and kernel parameters in the application of kernel principal component analysis is discussed. Based on the empirical formula of fitting information entropy and kernel function parameters based on UCI and statistical yearbook, the empirical formula 22291.11274-410.21300612.9958-xxx. Secondly, on the basis of summing up the objectives and principles of ecological efficiency index selection, aiming at the current domestic and foreign research on ecological efficiency and the existing problems, combined with the current economic development and the relationship between resources and environment, The original ecological efficiency evaluation and prediction index system were improved. Using kernel principal component analysis (KPCA) and data envelopment analysis (DEA), the classical ecological efficiency evaluation model was improved and the KPCA-DEA compound evaluation model was established. Then, the kernel principal component analysis (KPCA) method and BP neural network prediction method are used to improve the classical prediction model, and the KPCA-BP composite prediction model is established. The results of empirical analysis show that 14 of the comprehensive technical efficiency values of DEA model evaluation are 12 of the 1PCA-DEA models with a comprehensive technical efficiency value of 1. The synthetic technical efficiency of KPCA-DEA model is 6. The result shows that the composite evaluation model of KPCA-DEA is better than that of PCADEA and DEA. The empirical analysis of KPCA-BP composite prediction model shows that the relative error of KPCA-BP is 1.56and that of PCA-BP is 17.150.The relative error of BP model is 20.750.The result shows that KPCA-BP has a relative error of 20.75. the result shows that the relative error of KPCA-BP is 17.15, and that of BP model is 20.75. the result shows that the relative error of KPCA-BP is 17.15. The model has some advantages in prediction. This study enriches the evaluation and prediction methods of ecological efficiency and provides a reference for the evaluation and prediction of ecological efficiency.
【学位授予单位】：华北理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F205;O212.1

【参考文献】