准确预测电子商务销售额对电子商务业高质量发展具有重大意义。本文以贵州省为例,基于收集到的5个预测指标,建立可加模型对贵州省电子商务销售额未来三年进行预测。首先使用变分自编码器(VAE)方法对2003~2012年电子商务销售额缺失值进...准确预测电子商务销售额对电子商务业高质量发展具有重大意义。本文以贵州省为例,基于收集到的5个预测指标,建立可加模型对贵州省电子商务销售额未来三年进行预测。首先使用变分自编码器(VAE)方法对2003~2012年电子商务销售额缺失值进行估计补充,然后使用互信息评估5个预测指标被选择的合理性,同时使用基于B样条的核密度回归估计方法对可加模型进行估计。最后使用GM (1, 1)模型预测未来三年5个预测指标值,然后带入到已建立的可加模型中,从而得到电子商务销售额未来三年的预测值。Accurate prediction of e-commerce sales is of significant importance for the high-quality development of the e-commerce industry. Taking Guizhou Province as an example, this paper establishes an additive model to forecast the e-commerce sales in Guizhou Province for the next three years based on five collected predictive indicators. Firstly, the Variational Autoencoder (VAE) method is used to estimate and supplement the missing values of e-commerce sales from 2003 to 2012. Then, the mutual information is utilized to assess the rationality of the five predictive indicators selected, while the B-spline-based kernel density regression estimation method is employed to estimate the additive model. Finally, the GM (1, 1) model is used to predict the values of the five predictive indicators for the next three years, which are then incorporated into the established additive model to obtain the forecasted e-commerce sales for the next three years.展开更多
随着数字技术的不断发展和普及数字经济得到了迅速发展,对城市的发展带来了广阔的机遇。发展潜力分析对了解一个区域城市自身的产业发展规律和协助政府经济决策都具有相当重要的研究意义。我们可通过对比分析中国某省内在数字经济背景...随着数字技术的不断发展和普及数字经济得到了迅速发展,对城市的发展带来了广阔的机遇。发展潜力分析对了解一个区域城市自身的产业发展规律和协助政府经济决策都具有相当重要的研究意义。我们可通过对比分析中国某省内在数字经济背景下的其它各个主要城市及其经济发展及其潜力,做出哪些对其各大城市发展最为有利的区域发展的计划方案以及具体决策,从而使之发展得到进一步的提高。以贵州省9个市为例,采用主成分分析法,得出各市县的经济综合价值。利于综合制定产业发展与对策,以期能够为贵州省近期的发展经济战略进一步制定以及区域间协调发展提供参考。With the continuous development and popularization of digital technology, the digital economy has developed rapidly, bringing vast opportunities to the development of cities. The analysis of development potential is of great research significance for understanding the industrial development laws of a regional city itself and assisting the government in economic decision-making. We can compare and analyze the economic development and potential of other major cities in a certain province of China under the background of digital economy, and make specific plans and decisions that are most beneficial for the development of its major cities, so as to further improve their development. Taking nine cities in Guizhou Province as an example, using principal component analysis, we can obtain the comprehensive economic value of each city and county, which is conducive to the comprehensive formulation of industrial development and countermeasures. This is hoped to provide reference for the further formulation of economic development strategies and coordinated development between regions in Guizhou Province in the near future.展开更多
为求解黎曼流形上的大规模可分离问题,Kasai等人在(Advances of the neural information processing systems, 31, 2018)中提出了使用非精确梯度和非精确Hessian的黎曼信赖域算法,并给出了该算法的迭代复杂度(只有证明思路,没有具体证明...为求解黎曼流形上的大规模可分离问题,Kasai等人在(Advances of the neural information processing systems, 31, 2018)中提出了使用非精确梯度和非精确Hessian的黎曼信赖域算法,并给出了该算法的迭代复杂度(只有证明思路,没有具体证明)。我们指出在该文献的假设条件下,按照其思路不能证明出相应的结果。本文提出了不同的参数假设,并证明了算法具有类似的迭代复杂度。展开更多
文摘准确预测电子商务销售额对电子商务业高质量发展具有重大意义。本文以贵州省为例,基于收集到的5个预测指标,建立可加模型对贵州省电子商务销售额未来三年进行预测。首先使用变分自编码器(VAE)方法对2003~2012年电子商务销售额缺失值进行估计补充,然后使用互信息评估5个预测指标被选择的合理性,同时使用基于B样条的核密度回归估计方法对可加模型进行估计。最后使用GM (1, 1)模型预测未来三年5个预测指标值,然后带入到已建立的可加模型中,从而得到电子商务销售额未来三年的预测值。Accurate prediction of e-commerce sales is of significant importance for the high-quality development of the e-commerce industry. Taking Guizhou Province as an example, this paper establishes an additive model to forecast the e-commerce sales in Guizhou Province for the next three years based on five collected predictive indicators. Firstly, the Variational Autoencoder (VAE) method is used to estimate and supplement the missing values of e-commerce sales from 2003 to 2012. Then, the mutual information is utilized to assess the rationality of the five predictive indicators selected, while the B-spline-based kernel density regression estimation method is employed to estimate the additive model. Finally, the GM (1, 1) model is used to predict the values of the five predictive indicators for the next three years, which are then incorporated into the established additive model to obtain the forecasted e-commerce sales for the next three years.
文摘随着数字技术的不断发展和普及数字经济得到了迅速发展,对城市的发展带来了广阔的机遇。发展潜力分析对了解一个区域城市自身的产业发展规律和协助政府经济决策都具有相当重要的研究意义。我们可通过对比分析中国某省内在数字经济背景下的其它各个主要城市及其经济发展及其潜力,做出哪些对其各大城市发展最为有利的区域发展的计划方案以及具体决策,从而使之发展得到进一步的提高。以贵州省9个市为例,采用主成分分析法,得出各市县的经济综合价值。利于综合制定产业发展与对策,以期能够为贵州省近期的发展经济战略进一步制定以及区域间协调发展提供参考。With the continuous development and popularization of digital technology, the digital economy has developed rapidly, bringing vast opportunities to the development of cities. The analysis of development potential is of great research significance for understanding the industrial development laws of a regional city itself and assisting the government in economic decision-making. We can compare and analyze the economic development and potential of other major cities in a certain province of China under the background of digital economy, and make specific plans and decisions that are most beneficial for the development of its major cities, so as to further improve their development. Taking nine cities in Guizhou Province as an example, using principal component analysis, we can obtain the comprehensive economic value of each city and county, which is conducive to the comprehensive formulation of industrial development and countermeasures. This is hoped to provide reference for the further formulation of economic development strategies and coordinated development between regions in Guizhou Province in the near future.
文摘为求解黎曼流形上的大规模可分离问题,Kasai等人在(Advances of the neural information processing systems, 31, 2018)中提出了使用非精确梯度和非精确Hessian的黎曼信赖域算法,并给出了该算法的迭代复杂度(只有证明思路,没有具体证明)。我们指出在该文献的假设条件下,按照其思路不能证明出相应的结果。本文提出了不同的参数假设,并证明了算法具有类似的迭代复杂度。
