基于特征筛选算法的数字土壤制图研究

平缓地带数字土壤制图中，环境协变量的选择是提高制图精度的关键。已有研究证明遥感影像可作为推理制图的辅助因子，而如何确定环境因子推理制图时各自的权重已成为现阶段研究的重点。选取湖北省麻城市乘马岗镇为研究区，采用3种特征筛选方法进行有效环境变量筛选，探索参与平原-丘陵混合区域制图的因子并确定其重要性，依据选择的相对稳定的指标，进一步探索提高土壤类型制图准确性的途径。根据141个野外独立样点的检验结果表明：在推理制图中，遥感因子在平原区域的重要性程度高于丘陵区域，且遥感因子中归一化植被指数(NDVI)和均值(Mean)较为稳定；基于递归特征算法的按地形推理制图精度最高为75.89%，分别高于ReliefF算法和基于Tree的特征筛选算法13.48%和4.97%；此外3种特征筛选算法制图结果中，按地形因子分区制图的精度均高于整体区域制图。因此，遥感因子作为辅助手段参与推理过程可有效提高制图精度。本研究采用的特征挖掘与机器学习算法对提升土壤制图精度具有一定的理论意义。     

知识经济与洛阳经济可持续发展

知识经济时代，洛阳经济可持续发展战略应该是：加大教育科研改革，全面实施“科教兴洛”战略，夯实知识经济的基础产业；扩大农业的知识资本投入量，促进农业产业化、企业化、资本化经营，实现农业知识化和现代化；得用新技术成果，对传统工业进行知识化改造，着力营造一批高新技术产亚，推动洛阳产业结构全面升级和优化。     

棕榈在孟加拉乡村传统利用——一种可持续的生物多样性管理手段(英文)

采用多级采样的方法研究了孟加拉国农村耕作区内棕榈树的传统利用模式和当地的管理经验。农民管理棕榈主要为获取树液产品,生产以糖为主的次生品。树液可直接用作饮料,也可加工成糖蜜或酒精饮料。7块不同地块为棕榈提供生长环境,20.40%的棕榈种植在果园内。尽管中等级的农民拥有多数棕榈(33%),但是主要管理棕榈的是没有土地的农民,他们靠棕榈维持生计。这些农民将自己的智慧应用到棕榈管理当中,从种植到开孔取树液再到产品加工。如果能更科学地管理棕榈,加上农民的智慧,棕榈业将会对当地经济发展起到重要作用,也会丰富当地生物多样性。     

Towards the sustainable intensification of agriculture—a systems approach to policy formulation

The sustainable intensification of agriculture involves providing sufficient food and other ecosystem services without going beyond the limits of the earth’s system. Here a project management approach is suggested to help guide agricultural policy to deliver these objectives. The first step is to agree measurable outcomes, integrating formal policy goals with the often much less formal and much more diverse goals of individual farmers. The second step is to assess current performance. Ideally, this will involve the use of farm-scale metrics that can feed into process models that address social and environmental domains as well as production issues that can be benchmarked and upscaled to landscape and country. Some policy goals can be delivered by supporting ad hoc interventions, while others require the redesign of the farming system. A pipeline of research, knowledge and capacity building is needed to ensure the continuous increase in farm performance. System models can help prioritise policy interventions. Formal optimization of land use is only appropriate if the policy goals are clear, and the constraints understood. In practice, the best approach may depend on the scale of action that is required, and on the amount of resource and infrastructure available to generate, implement and manage policy.     

Renaturing cities: from utopias to contested realities and futures

Renaturing urban environments is a transformative pathway for urban sustainability that can be leveraged for collaborative research and planning to reverse long trends of ecosystem degradation. People-nature connections need to be reinforced to enable the successful uptake and upscale of urban renaturing practices. Improving people’s understanding, perception, and emotions towards nature is therefore key. In this paper, we discuss how human knowledge and values of nature can be enabled through urban renaturing. Besides, we discuss the required transitions in urban planning processes to support urban renaturing practices.     

基于CiteSpace的氮沉降知识图谱计量分析

客观分析国内氮沉降研究现状,总结现阶段研究的重点与薄弱点,旨在为氮沉降的科研工作者与决策者提供数据参考.以"氮沉降"为主题词,从CNKI数据库中得到2009-2019年发表的相关文献754篇,运用CiteSpace知识图谱分析工具对关键词、作者、机构进行可视化分析,并对氮沉降的研究进展、发展趋势和资助基金等方面进行相关...     

A Vovel Constracture of Combinatorial Knowledge Base for the Expert System of Eletrical Switching

By comparing the techniges of the production and the logic predicate for knowledge representation, this paper presents a kind of constructure of combinatorial knowledge base applied to the expert system for switch operation in electric power system . This constructure of knowledge base to be expanded and revised easily,and it is cinvenitnt to build the general shell of the expert system.     

基于知识模型的棉花管理决策支持系统

通过对棉花栽培研究资料和专家知识的总结归纳,运用系统学方法和结构化途径,建立了基于知识模型的棉花管理决策支持系统(KMDSSCM)。它由数据库、知识模型库、专家咨询库和人机界面等部分组成;可以实现栽培方案设计、适宜生育指标确定、适时苗情调控、专家知识咨询和系统维护等功能,具有普适性强,独立性好等特点。实例分析结果表明,系统可以根据决策点的基础农情条件,设计出适宜的综合性栽培管理方案和生育指标动态。     

多邻域保持嵌入的人脸识别方法

现有流形学习算法在学习人脸数据时,假设所有数据点位于单一低维嵌入流形之上,当数据点实际分布在不同的流形上时,单流形假设就会影响数据真实空间结构。为此提出一种基于多邻域保持嵌入（multiple neighborhood preserving embedding, M-NPE）的学习算法来发现不同类别数据在不同维度的低维嵌入空间中分布的多流形结构。首先,单独学习不同类别数据的流形,得到反映其本质特征的流形;再通过遗传算法搜索每个流形的最优维数;最后依据最小重构误差分类器对样本分类。在Extended Yale B和CMU PIE这2个大型人脸库上实验结果验证了该算法的有效性。     

A cropping system assessment framework—Evaluating effects of introducing legumes into crop rotations

Methods are needed for the design and evaluation of cropping systems, in order to test the effects of introducing or reintroducing crops into rotations. The interaction of legumes with other crops (rotational effects) requires an assessment at the cropping system scale. The objective of this work is to introduce a cropping system framework to assess the impacts of changes in cropping systems in a participatory approach with experts, i.e., the integration of legumes into crop rotations and to demonstrate its application in two case studies. The framework consists of a rule-based rotation generator and a set of algorithms to calculate impact indicators. It follows a three-step approach: (i) generate rotations, (ii) evaluate crop production activities using environmental, economic and phytosanitary indicators, and (iii) design cropping systems and assess their impacts. Experienced agronomists and environmental scientists were involved at several stages of the framework development and testing in order to ensure the practicability of designed cropping systems. The framework was tested in Västra Götaland (Sweden) and Brandenburg (Germany) by comparing cropping systems with and without legumes. In both case studies, cropping systems with legumes reduced nitrous oxide emissions with comparable or slightly lower nitrate-N leaching, and had positive phytosanitary effects. In arable systems with grain legumes, gross margins were lower than in cropping systems without legumes despite taking pre-crop effects into account. Forage cropping systems with legumes had higher or equivalent gross margins and at the same time higher environmental benefits than cropping systems without legumes. The framework supports agronomists to design sustainable legume-supported cropping systems and to assess their impacts.