二次曲线拟合算法的统计性能分析与改进

传统的二次曲线拟合使用标准特征值分析算法。通过统计分析技术 ,可知该技术在拟合数字二次曲线时 ,存在估计偏差大、均方误差大的缺点。其产生原因是数据噪声的有色性和自相关函数矩阵的条件数过大 ,因此白化数据噪声和正则化变换是提高曲线拟合的有效措施。这从理论上有力地支持了Hartley提出的正则化算法。通过理论分析和计算机仿真实验 ,表明了降维EVD技术固有地同时具备噪声预白化功能和数据正则化功能 ,因此它能给出均方误差相当小的无偏估计。由于它无须进行预白化变换或正则化变换 ,并把最优化过程的维数从 6降为 2 ,所以它还具有计算快速、实现简单方便的优点。     

Constructivisms and Relativisms: A Shopper‘s Guide

Diverse forms of constructivism can be found in the literature today. They exhibit a commonality regarding certain classical positions that they oppose - a unity in their negative identities - but a sometimes wild multiplicity and incompatibility regarding the positive proposals that they put forward. In particular, some constructivisms propose an epistemological idealism, with a concomitant relativism, while others are explicitly opposed to such positions, and move in multifarious different directions. This is a potentially confusing situation, and has resulted in some critics branding all constructivisms with the charge of relativism, and throwing out the baby with the bath water. In addition, since the epistemological foundations of even non-relativist constructivisms are not as familiar as the classical positions, there is a risk of mis-interpretation of constructivisms and their consequences, even by some who endorse them, not to mention those who criticize. Because I urge that some version of constructivism is an epistemological necessity, this situation strikes me as seriously unfortunate for philosophy, and potentially dangerous for the practice of education.     

表象训练法在背越式跳高教学中的应用

将表象训练的原理运用于背越式跳高技术的教学过程中,为改进传统的教学方法,提高教学质量进行了有益的探讨和尝试.     

体育科研三维方法述要(上)

将常规的体育科研方法区分为三个相互关联的维度，其中，取材维含括观察、实验、假设三层阶，分别强调思考、目标、理论的不可或缺性；程序维依循一般思辨、专一选题、采集素材、整理素材、推理判断，验察修正六环节；课题维则从生物性学科和社会性学科两大方面寻求变革。最后还论及撰写文章十要点和缩减文字三要诀。     

基于无监督图神经网络的学术文献表示学习研究

学术文献特征表示,是学术文献搜索、分类组织、个性化推荐等学术大数据服务的关键步骤。研究表明,图神经网络能够有效学习文献的特征表示,然而当前研究主要集中在有监督学习方法上,不仅对数据集的大小和质量的要求较高,且学习到的文献特征表示与具体任务高度耦合。基于此,本文将四种无监督图神经网络方法引入学术文献表示学习,从Cora、CiteSeer和DBLP (database systems and logic programming)数据集的引文网络、共被引网络和文献耦合网络中学习文献的表示向量,并应用于文献分类和论文推荐两大下游任务。研究结果表明,(1)深度互信息图神经网络适合于文献分类任务,对抗正则化变分图自编码器则在论文推荐任务上性能更佳;(2)Cora数据集上的结果表明,相较于共被引和文献耦合网络,引文网络更适合于学习通用的文献表示向量。     

质性知识构造与量化知识分析综论:兼论学术对标法理论基础

文章在简要综述知识基因、知识单元、知识表示和知识图谱的基础上,探讨知识的质性构造和量化分析基础,指出知识具有表型(外型)、构型(内型)和量型(量值)三型特征。用逻辑分析法从现象到本质揭示知识表型特征,包括语言文字型、公式符号型和图表图谱型;知识构型可用学科类别和特征概念标识;知识量型可基于数据或信息计算知识量。进一步提出用知识花图式和知识花程式展示质性知识构造与量化知识分析。由知识三型特征引出知识变换三假设以及学术评价"对标法";揭示质性知识构造适合解析静态知识特征,量化知识分析可延伸处理动态知识演化。质性知识构造与量化知识分析同时适用于剖析知识客体和知识主体,并独具学术评价应用价值。倡导通过知识分析创新图书情报学,同时也通过图书情报学促进知识之花盛开。     

信息检索可视化基本问题研究

信息检索可视化的基本问题有:信息检索模型、信息内容描述、可视化映射技术、可视化显示技术、全局映射与局部映射、实时可视化和人工参与的可视化等。可根据实际情况合理选择,可选择一种,也可把多种技术组合在一起。表1。参考文献21。     

Inverted Index Compression Using Word-Aligned Binary Codes

We examine index representation techniques for document-based inverted files, and present a mechanism for compressing them using word-aligned binary codes. The new approach allows extremely fast decoding of inverted lists during query processing, while providing compression rates better than other high-throughput representations. Results are given for several large text collections in support of these claims, both for compression effectiveness and query efficiency.     

政府网站信息资源知识元模型与可视化表征研究

[目的/意义]利用知识元模型理论研究政府网站知识服务效果的优化路径,辅以可视化表征技术,以降低大数据环境下政务用户信息获取的操作负载和知识加工的认知负荷。[方法/过程]依据相关知识元模型研究推理出符合政府网站信息资源属性特征的六元组知识元表示方法和四元组知识元本体结构,采用TextRank与HDP算法分别抽取政府网站信息资源关键词和主题词,并由领域专家根据抽取结果确定知识元,构建包含知识元本体库生成和可视化知识服务的政府网站信息资源领域知识元可视化表征模型。[结果/结论]通过政府网站发布的共享单车实例检验知识元可视化表征模型的有效性和可行性,为实现政府网站粗粒度信息服务转向以知识元为单位的细粒度知识服务范式开辟了新的研究思路,可视化知识服务模式增强了政务信息导航的结构化和用户解读领域文本语义的效果。     

Proximal advantage in knowledge diffusion: The time dimension

This paper intends to explore the impact of geographic proximity on the diffusion of knowledge in the form of publication citations, and argues that codified knowledge is transmitted faster in proximity and is subject to similar geographic constraints as tacit knowledge. The geographic proximity advantage would be particularly relevant in the early stage of dissemination. We collected three sets of research articles published in 1990, 2000 and 2010 and compared the longitudinal citations they received domestically and from abroad. The study found that domestic citations accumulate faster and reach their peak much earlier than foreign citations, and the difference is most evident in the first few years after publication. The result shows that geographic proximity does play a role in the speed of knowledge diffusion and points to the network effect for citations. Those located closer to the knowledge origin would be exposed and react to publications faster due to the additional opportunities of research exchange and network.