为了能够更加准确地实现地铁客流预测,提出了一种基于经验模态分解算法(empirical mode decomposition,EMD)优化非线性自回归(nonlinear auto regressive,NAR)动态神经网络的地铁客流量短时预测模型.分析地铁客流量数据后发现日客流量...为了能够更加准确地实现地铁客流预测,提出了一种基于经验模态分解算法(empirical mode decomposition,EMD)优化非线性自回归(nonlinear auto regressive,NAR)动态神经网络的地铁客流量短时预测模型.分析地铁客流量数据后发现日客流量具有一定的变化规律,为此使用了基于时间序列的NAR动态神经网络,该网络具有优秀的非线性动态拟合能力和反馈记忆的功能.结合EMD经验模态分解算法优化NAR动态神经网络预测模型,以此来减少预测误差,提高预测精度.结果显示,EMD-NAR神经网络组合预测模型适用于地铁客流的短时预测,预测精度可达93%,具有较好的应用价值.展开更多
This paper proposes two concepts: the ecological footprint component index(EFCI) and the biocapacity component index(BCCI), based on the ecological footprint(EF) and Shannon entropy approaches. Per capita EFCI and BCC...This paper proposes two concepts: the ecological footprint component index(EFCI) and the biocapacity component index(BCCI), based on the ecological footprint(EF) and Shannon entropy approaches. Per capita EFCI and BCCI in China 1949-2013 are analyzed using empirical mode decomposition(EMD). Nonlinear models of per capita EFCI and BCCI in China 1949-2013 are presented and their cycles and predictions from 2014 to 2023 are analyzed. The results over the last 65 years show:(1) EFCI in China has increased constantly with fluctuations, while BCCI has slowly decreased. Their annual change rates are 2.81% and-1.26%, respectively. The increasing EFCI indicates a gradual improvement in China's sustainable development potential; the decreasing BCCI indicates severe environmental and population challenges.(2) The cycles of per capita EFCI have periods of 5.4 and 16.3 years, while cycles of per capita BCCI have periods of 3.6, 13,and 21.7 years. The predictive models indicate that EFCI will first decrease, reaching 0.02725 in2014, and will subsequently increase to 0.03261 in 2021. BCCI will increase, reaching 0.01365 in2014 and 0.01541 in 2022. EFCI and BCCI will reach 0.03037 and 0.01537, respectively, in 2023.Policymakers should ensure that the EFCI and BCCI increase in 2023.展开更多
文摘为了能够更加准确地实现地铁客流预测,提出了一种基于经验模态分解算法(empirical mode decomposition,EMD)优化非线性自回归(nonlinear auto regressive,NAR)动态神经网络的地铁客流量短时预测模型.分析地铁客流量数据后发现日客流量具有一定的变化规律,为此使用了基于时间序列的NAR动态神经网络,该网络具有优秀的非线性动态拟合能力和反馈记忆的功能.结合EMD经验模态分解算法优化NAR动态神经网络预测模型,以此来减少预测误差,提高预测精度.结果显示,EMD-NAR神经网络组合预测模型适用于地铁客流的短时预测,预测精度可达93%,具有较好的应用价值.
基金supported by the Opening Foundation of Jiangsu Key Laboratory of Environment Change&Ecological ConstructionNational Natural Science Foundation of China:[Grant Number 41372182]Research Center of Resource-exhausted Cities Transformation and Development:[Grant Number Kf2013y08]
文摘This paper proposes two concepts: the ecological footprint component index(EFCI) and the biocapacity component index(BCCI), based on the ecological footprint(EF) and Shannon entropy approaches. Per capita EFCI and BCCI in China 1949-2013 are analyzed using empirical mode decomposition(EMD). Nonlinear models of per capita EFCI and BCCI in China 1949-2013 are presented and their cycles and predictions from 2014 to 2023 are analyzed. The results over the last 65 years show:(1) EFCI in China has increased constantly with fluctuations, while BCCI has slowly decreased. Their annual change rates are 2.81% and-1.26%, respectively. The increasing EFCI indicates a gradual improvement in China's sustainable development potential; the decreasing BCCI indicates severe environmental and population challenges.(2) The cycles of per capita EFCI have periods of 5.4 and 16.3 years, while cycles of per capita BCCI have periods of 3.6, 13,and 21.7 years. The predictive models indicate that EFCI will first decrease, reaching 0.02725 in2014, and will subsequently increase to 0.03261 in 2021. BCCI will increase, reaching 0.01365 in2014 and 0.01541 in 2022. EFCI and BCCI will reach 0.03037 and 0.01537, respectively, in 2023.Policymakers should ensure that the EFCI and BCCI increase in 2023.
