摘要
传统DEA方法存在小样本估计有偏、无法进行统计检验等缺点.而Bootstrap-DEA方法能够通过数值模拟技术产生大量的模拟样本,通过对新生成的样本进行估计来修正DEA效率估计偏差,进而计算出效率值的置信区间.作者利用Bootstrap-DEA方法首次对我国36个工业行业1995-2008年的能源利用效率进行分析.研究结果表明,我国工业行业的能源效率呈现出"U"字型变动趋势,行业间的能源效率差异不断减小.
The traditional DEA mothod has some shortcomings such as small sample＇s estimation exists biase and statistics test can＇t be conducted,while Bootstrap-DEA method can create many simulated samples by numerical simulation and fix the bias of energy efficiency estimated by DEA method,and then calculate the energy efficiency＇s confidence interval.In this paper,we first employ the Bootstrap-DEA method and analyze the energy efficiency of China＇s 36 industries from 1995 to 2008.The research results show that the industries energy efficiency changes with ＂U＂ shape,and the gap among the industries efficiency decreases continuously.
出处
《系统科学与数学》
CSCD
北大核心
2011年第3期361-371,共11页
Journal of Systems Science and Mathematical Sciences
基金
中国科学院重大方向性项目"全球经济监测预警及政策模拟仿真平台预研项目"(编号KACX1-YW-0906).
