基于卡尔曼滤波算法的焦炭价格预测（英文）

焦炭价格预测研究具有重要的理论和实践意义,本文利用卡尔曼滤波算法对焦炭价格进行预测研究。建立状态空间模型时,选取焦炭价格作为唯一的状态变量,通过每一时刻变量观测值与预测值形成的新息,不断更新和迭代,以寻求最优估测值。实证分析表明,该算法对焦炭价格的跟踪和预测效果较好。     

用卡尔曼滤波算法预测埃塞俄比亚咖啡价格（英文）

埃塞俄比亚咖啡价格波动很大,因此对国家经济发展的影响不容小视,对咖啡价格进行预测具有理论和实践意义。为了分析咖啡价格波动,我们采用来自埃塞俄比亚商品交易所(ECX)记录的2008年6月25日至2017年1月5日期间咖啡日收盘价数据。在这里,咖啡价格的性质是非平稳的,我们在单个线性状态空间模型上应用卡尔曼滤波算法来预测咖啡价格的最优值,主要通过使用均方根误差(RMSE)来评估用于预测咖啡价格的算法的性能。基于线性状态空间模型和卡尔曼滤波算法,均方根误差(RMSE)为0.000016375,说明该算法性能良好,研究结果可靠。     

基于单位汇水面积的地形因子特征评价

探讨了典型黑土区基于单位汇水面积的LS因子算法与传统基于径流路径长度算法的差异。研究表明,在研究区坡长坡缓的地形条件下,基于单位汇水面积的LS因子算法值普遍要大于基于径流路径长度算法的LS因子值;在有限的坡度范围内,不同算法的坡度因子相差不大,与之相比,不同算法的坡长因子值差别明显,特别是对于坡长指数采用定值的Govers算法更是如此。这说明在研究区坡长坡缓的地形条件下,坡长因子对LS因子算法的响应更为敏感。同时,侵蚀机理分析得出,坡长因子指数采用与坡度有关的变值更为合理,特别是在采用DEM进行大范围土壤侵蚀量预测时更应如此。     

卧虎山水库不同算法LS因子值适用性分析北大核心CSCD

在总结以往LS因子相关算法研究基础上,结合Mc Cool的LS因子参照值,对LS算法的适用性进行评价分析。研究表明,除去复合算法和Remortel修正算法外,其他不同算法LS因子值都小于参照值;研究区LS因子值的最优算法为基于Remortel迭代运算的修正算法;其次为结合刘宝元的陡坡公式和Remortel改进L指数因子迭代运算的复合算法,以及Remortel第4版AML程序算法;再次为B?hner算法,而Moore算法和Desmet算法,由于其与参照值的相关性相对较差,而且其RMSE相对较大,不推荐在该区使用。     

我国钢铁行业下半年发展趋势展望

2011年对于我国钢铁行业来说是机遇与挑战并存。由于房地产、制造业需求增速在上半年纷纷出现下滑趋势,对钢材消费的拉动作用明显不足。铁矿石、焦炭等主要原料价格暴涨直接     

卧虎山水库不同算法LS因子值适用性分析

在总结以往LS因子相关算法研究基础上,结合McCool的LS因子参照值,对LS算法的适用性进行评价分析。研究表明,除去复合算法和Remortel修正算法外,其他不同算法LS因子值都小于参照值;研究区LS因子值的最优算法为基于Remortel迭代运算的修正算法;其次为结合刘宝元的陡坡公式和Remortel改进L指数因子迭代运算的复合算法,以及Remortel第4版AML程序算法;再次为B?hner算法,而Moore算法和Desmet算法,由于其与参照值的相关性相对较差,而且其RMSE相对较大,不推荐在该区使用。     

基于残差自回归与Kalman滤波的天然气消费量组合预测研究（英文）

目前中国以化石能源为主的能源消费格局正逐步优化,清洁能源的消费规模逐步增加。对天然气消费量的预测分析对未来能源消费结构调整具有积极意义。本文创新性地采用拟合值偏离度倒数法进行权重设置,利用残差自回归模型和Kalman滤波算法构建组合预测模型,以《BP世界能源统计年鉴》和《中国统计年鉴》1980–2017年的天然气消费历史数据为对象,对中国天然气消费量进行预测研究。研究结果表明:(1)组合预测模型的预测精度更高:残差自回归预测模型的相对误差落在(–0.08,0.09)区间内,卡尔曼滤波预测的相对误差落在(–0.09, 0.32)区间内,组合预测模型相对误差落在(–0.03, 0.11)区间内。(2)组合预测模型预测结果的稳定性更好:残差自回归预测模型相对误差的预测方差为0.002,卡尔曼滤波预测相对误差的预测方差为0.007,组合预测模型相对误差的预测方差为0.001。(3)其他条件不变的情况下,2018年天然气消费量费预测值为2418.08亿m~3。与其他时间序列预测方法相比,利用残差自回归模型和卡尔曼滤波算法构成的组合预测模型对数据限制条件少,可操作性强,且分析结果更为可信。     

基于可调功效函数容量耦合生态经济协调发展评价模型

依据功效函数和熵值赋权法的原理,改进min-max功效函数和熵值赋权法的计算基础,提出可调可比可预测的功效函数与基于均值功效的熵值赋权法;讨论生态经济协调发展的等级标准,阐明容量耦合系数模型与生态经济协调发展的评价算法;以精简评价指标和鄱阳湖生态经济建设战略思想为指导,建立生态经济协调发展评价指标体系。将生态经济协调发展评价的等级标准、模型算法和指标体系在江西省九江市庐山区加以实际应用,验证模型的适用性。     

中国城市住宅价格时空演变研究进展与展望

城市住宅价格时空演变历来是城市地理学和城市经济学关注的核心内容之一。住宅价格在空间、时间分布上具有内在规律性,探索这一规律信息及其影响机制,有助于政府制定合理的房价调控和土地利用政策。本文在总结中国城市住宅价格时空演变动力机制与模式、影响因素以及技术方法等方面研究进展的基础上,指出在研究内容与研究方法方面仍存在不足,有待深化。研究结论如下:未来中国城市住宅价格时空演变研究应立足于经济转型与制度改革的特殊背景,重视住宅价格时空演变理论框架的构建;加强宏观因素对住宅价格影响的定量研究以及探索住宅价格低值集聚的内外因素,全面考察各类因素对住宅价格的影响;强化GIS 技术对城市住宅价格时空演变特征信息的挖掘,并探索多维空间在时间序列上的动态GIS 模型,实现对住宅价格时空演变与分布的模拟分析与趋势预测;拓展城市住宅价格时空演变的普及性应用,并将研究成果与城市住宅价格监测成果形成动态互补性应用。     

抚州市城镇住宅价格影响因子相关性研究

针对抚州市城镇住宅价格空间的演变与时序变化,利用相关分析法、熵值法和主成分分析法确定稳定性评价指标,把基于空间特征的城镇住宅价格影响因子用于相关性研究中,对城镇住宅价格影响因子的选取、影响因子的相关性、权重的确定等问题进行研究。结论：经济状况因子是城市商品住宅价格的决定因子,社会状况因子是影响商品住宅价格的重要因子,制度因素因子是影响城市商品住宅价格的一个不可忽略的方面。     

Forecasting the Ethiopian Coffee Price Using Kalman Filtering Algorithm

Ethiopian coffee price is highly fluctuated and has significant effect on the economy of the country. Conducting a research on forecasting coffee price has theoretical and practical importance.This study aims at forecasting the coffee price in Ethiopia. We used daily closed price data of Ethiopian coffee recorded in the period 25 June 2008 to 5 January 2017 obtained from Ethiopia commodity exchange (ECX) market to analyse coffee prices fluctuation. Here, the nature of coffee price is non-stationary and we apply the Kalman filtering algorithm on a single linear state space model to estimate and forecast an optimal value of coffee price. The performance of the algorithm for estimating and forecasting the coffee price is evaluated by using root mean square error (RMSE). Based on the linear state space model and the Kalman filtering algorithm, the root mean square error (RMSE) is 0.000016375, which is small enough, and it indicates that the algorithm performs well.     

Accuracy of forecasting in a Nordic paper mill's supply chain: A case study

The article compares the accuracy of forecasted and actual supply chain cost and price factors in a Nordic paper mill's sales to customers in two large European countries. Empirical longitudinal research data covering the years between 2002 and 2008 were obtained from a large integrated Finnish paper mill, and consisted of sales volumes, paper prices, variable costs, and transport costs. It was expected that the mill would be able to forecast demand, prices and costs accurately, but the empirical findings showed that the forecasts for paper price, demand, and cost varied from one market to another and were not highly accurate. In addition, the forecasting of gross margins seemed to be very inexact. It is concluded that there is a need for more reliable forecasting methods in the paper industry to anticipate economic development as paper demand and costs change.     

Forecasting Copper Prices Using Hybrid Adaptive Neuro-Fuzzy Inference System and Genetic Algorithms

Natural Resources Research - An accurate forecasting model for the price volatility of minerals plays a vital role in future investments and decisions for mining projects and related companies. In...     

Improving Adaptive Neuro-Fuzzy Inference System Based on a Modified Salp Swarm Algorithm Using Genetic Algorithm to Forecast Crude Oil Price

Natural Resources Research - This paper presents a novel forecasting model for crude oil price which has the largest effect on economies and countries. The proposed method depends on improving the...     

Innovation Clusters and Prospects for Environmental Management in Russia

We examine the approach to studying and forecasting promising markets of rational nature management. The relationships between higher educational institutions—centers of scientific research competences with other organizations are determined. It is found that the flow of knowledge from universities must generate stable regional networks (innovation clusters) to concentrate a significant part of markets. It is established that the most advanced innovation clusters emerged in regions with centers of potential interregional clusters in Moscow, St. Petersburg, Novosibirsk and Tomsk as well as in Perm krai and Tyumen, Belgorod and Kaliningrad oblasts. Largely on the basis of data on funding of state programs, we estimated the total potential volume of the market until 2020. For estimating the proportion of the regions in it, it is suggested that it will be higher if the organizations of relevant regional cluster show a higher publication and patent activity with a large number of interactions. Moscow will concentrate more than 20% and St. Petersburg about 10% of the market of rational environmental management. The total volume of national products and services will make up from 0.2 to 2% of the world market of the sphere under investigation by the year 2020; therefore, it is necessary to enhance the interactions of innovation organizations within the framework of cluster initiatives. But the main problem involves the absence of international cooperation. In fact, these efforts constitute one of the first attempts to study the sector associated with applied research of Earth sciences in Russia.     

BP网络模型在沙尘暴预测中的应用研究

依据锡林郭勒地区30 a气象资料,应用人工神经网络(ANN)中不同BP网络结构和算法,探索建立沙尘暴预测模型的方法。研究认为,在建立锡林郭勒地区沙尘暴预测模型时,选择年大风日数、年平均地温、年蒸发量、相对湿度4个气象因子作为模型的输入因子是合理的;经过输入因子确定,层数、节点选择,每层激活函数和输出因子的确定,锡林郭勒地区沙尘暴预测模型可采用三层网络结构（4-6-1）。比较和试算显示,快速BP算法较普通BP算法的训练速度快,收敛精度高64.47%;快速BP神经网络的沙尘暴预测模型的预测精度可达到98%,较传统的多元线性回归数学模型高。因此,应用快速BP神经网络建立沙尘暴预测模型简捷、方便,具有精度高、智能化等特点,可在区域沙尘暴预测预报领域推广。     

State‐Led Innovation at the City Level: Policy Measures to Promote New Energy Vehicles in Shenzhen,China

Unlike many Western countries, the Chinese state has a much more active role in innovation that includes direct involvement with actors in innovation processes and the use of state capacity, money, and power to attain certain goals. With the rise of China, innovation processes that critically depend on state action, often labeled state‐led innovation, have thus received growing scholarly attention. However, only few contributions study the specific sector of new energy vehicles (NEV) and policy measures applied at the city level. This paper examines policy and planning tools used in Shenzhen, China, to assert innovation in the NEV industry, using an evaluation of documents and interviews. The paper finds that a city can play a decisive role in the implementation of innovation policy, and Chinese cities in particular make use of a broad set of innovation support measures ranging from binding quotas, public procurement, and incentives to bans and orders. The findings underline the importance of strong regulatory instruments that do not conform to the Western notion of market‐compliant policy, but nevertheless work effectively in the Chinese context. Moreover, the results highlight how successful policy support for innovation in the NEV can be implemented.     

基于粒子群双重空间聚类的城镇地价空间分异规律挖掘

城镇地价的空间分异对其动态监测片区划定具有重要的参考作用,而空间聚类则是对该问题进行分析的一种有效方法.文中在分析空间聚类本质的基础上,设计了一种基于样本数据的粒子群双重空间聚类挖掘模型,并以北海市二手房地价空间分布挖掘为实例进行了应用.结果表明:粒子群双重空间聚类在空间聚类效果上优于传统K-均值聚类法,尤其是对于任意形状聚类簇的挖掘;然而,由于标准粒子群优化算法机理的限制,其聚类效果随聚类中心的增多而呈下降趋势.总体而言,该方法能利用有限的样本挖掘质量较好的城镇地价空间分异区,可为城镇地价动态监管提供相     

农业产业集群创新效率及影响因素——基于山东省寿光蔬菜产业集群的实证分析

创新是当前农业产业集群研究的重要内容,掌握集群创新状况对指导农业产业集群发展具有重要意义。结合实地调研的数据,利用DEA-Tobit两阶段模型,对山东省寿光蔬菜产业集群企业创新效率及其影响因素进行研究。结果表明:寿光蔬菜产业集群规模已接近最优状态,综合创新效率处于中等偏上水平,但仍未达到最优状态,主要原因是集群在科研技术投入、创新资源利用配置等方面不足造成的。进一步Tobit回归的影响因素分析发现:企业知识技术吸收扩散能力、政府财政支持力度、企业家创新精神和企业发展的外向性与农业产业集群企业创新效率存在显著的正相关关系,而企业规模对农业产业集群企业创新效率影响并不显著。最后,根据集群创新影响因素分析,提出发展建议。     

长江中游城市群创新平台空间分布及其影响因素分析

创新是引领发展的第一动力。创新平台作为区域创新体系的重要组成部分和集聚创新资源的核心支撑载体,是推动创新驱动发展的主要抓手。论文以长江中游城市群为研究区域,运用最邻近指数、核密度估计、Ripley's K函数、地理探测器等方法,探究2017年长江中游城市群省级以上创新平台的空间分布特征及其影响因素。结果表明：① 长江中游城市群创新平台整体呈现出集聚分布特征,以武汉、长沙和南昌为核心呈“品字形”的分布格局;② 基于不同等级和不同功能类型划分的各类创新平台均呈现出显著集聚的特征,但集聚强度和集聚状态存在一定差异;③ 创新平台总体和不同类型创新平台的空间集聚存在尺度效应,随地理距离的变化空间集聚趋势先增强后减弱;④ 创新平台的空间分异是多因素综合作用的结果,金融实力、信息化程度、利用外资水平、人力资本条件的影响最为显著,其次为交通可达性、城镇化水平、经济基础和创新氛围。