With the development of China’s economy, environmental pollution has become cumulatively serious. The primary source of environmental pollution is thermal power generation, which has attracted the attention of governments and academia in recent years. To effectively reduce environmental pollution, research should study how to constrain the undesirable output of thermal power plants, that is, to limit the total undesirable output of the plants to a certain fixed sum. However, few studies have suggested that these undesirable outputs should be fixed-sum outputs. Moreover, no previous research publication about thermal power plants has analyzed their environmental performance changes. To address these gaps, a novel Malmquist-DEA approach is proposed for evaluate the environmental performance of thermal power plants in this paper. This approach generalizes the equilibrium efficient frontier DEA model with fixed-sum undesirable outputs and incorporates the model into the Malmquist productivity index (MPI). The authors apply this approach to the analysis of provincial thermal power plant environmental performance in China and analyze such plants’ trends based on panel data from 2011 to 2014. The empirical research shows that the environmental performance of regional thermal power plants was positively affected by efficiency change and negatively affected by technical change. Finally, the authors provide policy suggestions based on our findings.
现有结合特征提取与预测模型的方法不能准确把握金融时间序列的混沌性与交互性,导致预测精度不高。针对此问题,提出一种基于二次分解与长短期记忆(long short term memory, LSTM)网络的金融时间序列预测算法。使用变分模态分解方法与集成经验模态分解方法依次解析金融时间序列数据,得到能表达数据混沌性特征的模态;将模态信息输入到融合有因子分解机(factorization machine, FM)的长短期记忆网络模型中,融合获取到的长记忆性特征与交互性特征,进而预测最终的结果;选取沪深300指数的历史数据作为实验数据集,通过多组对比实验验证算法的有效性。实验结果表明,提出的算法可以有效提升模型的预测能力,同时表达金融时间序列的混沌性、长记忆性、交互性。 相似文献
A laboratory method combining the high-energy mechanical ball milling and high static pressure has been suggested for modelling synthesis of coesite in the Earth's crust. A window of milling time, a mechanical collision-induced intermediate phase ofα-quartz and its condition of easily crystallizing into coesite under the condition of 3.0 GPa, 923 K, < 1.0 min have been discovered. The new method needs much shorter synthesizing time and lower synthesizing critical pressure than those used before. The Raman spectra for the coesite synthesized by the present method have covered all the information of those natural and synthesized coesites reported before. It is demonstrated that this method has a great potential in geo-science research and implied the other possible mechanism of coesite formation in the Earth's crust. 相似文献