理性人的人情味：快手“老铁经济”的内在机制和逻辑

以中国用户规模最大短视频平台之一快手头部主播直播带货推广项目的案例和数据分析为基础,构建解释快手直播电商经济底层逻辑的社会学机制模型.提出快手直播电商经济背后的3种社会学机制:关键影响者逻辑、身份认同逻辑和社会网络逻辑.通过LDA文档主题生成模型将快手主播的直播语义字段进行文档主题生成归类,发现情感因素和理性因素词交互影响受众的购买决策,它背后的社会学作用机制是关键影响者逻辑和身份认同逻辑.2020年快手直播电商半年交易额已达1044亿元,理解快手直播背后的影响机制,能够持续优化推广模型,在互联网流量广告领域具有重要的应用价值.     

Multilayer financial networks and systemic importance: Evidence from China

In this paper, we develop a multilayer network structure and reveal the relationship between network structure and systemic risk. Unlike many previous studies, our model considers both liability and cross-holding of shares between financial institutions simultaneously. We propose a new systemic risk measurement by exploring the dynamic mechanism of financial contagion in the multilayer network. We display the network structure of Chinese financial institutions, including connectivity and diversity, and identify the systemic importance of them. We demonstrate that the multilayer network plays a non-linear role in financial risk spreading. Using the panel regression model and several experiment evidences, we show that the systemic risk can be explained more effectively by the linkage diversity more than the connectivity at both the institutional level and the system level. Our results highlight the importance of considering contagion mechanisms that go beyond a simple single-layer network structure.     

Application of collective knowledge diffusion in a social network environment

ABSTRACT

We present a group dynamics model that shows knowledge integration as a process occurring over time. As each individual in the group contact others, his own knowledge changes, and over time the collective knowledge is obtained. This allows modeling knowledge diffusion in a social network and while the models presented in this paper are not competitive in that area, they approach the problem from previously unconsidered direction. We test the behavior of the model in a multi-agent simulation and we test a simple advertisement campaign in a social network. We provide discussion of elements needed for making model more competitive.     

Market microstructure in a global B2B network

On‐line marketplaces raise several interesting issues, among them the relevance of location when content is digitized, and the assessment of a supplier's capabilities when buyers worldwide only have electronic contact with sellers. In global B2B on‐line marketplaces, market microstructures, i.e. which firms compete for the same customers, are thus likely to be influenced by how customers value location and firm capabilities in their decisions to do business with different suppliers on‐line. We suggest that both these sets of attributes will continue to matter on‐line—firms possessing similar capabilities, as well as firms that are similar in location by country, time zones or clusters, will compete for business from the same customers. We model the similarity in competitive positions between pairs of firms based on the overlap in their customer networks, using data on actual interactions between supplier and customer banks on an electronic trading system. Using QAP network regression techniques on the 100 largest banks in this industry, we find that similarity in capabilities influences who competes with whom, and that location still matters in a global B2B exchange. Interestingly, location influences who a firm's competitors are, but not where its customers are from. Copyright © 2001 John Wiley & Sons, Ltd.     

基于Elman神经网络的商业银行信贷风险识别及评估模型

建立了基于Elman神经网络的商业银行信贷风险识别及评估模型，并通过实例验证了模型的准确性和可靠性。研究过程及结果表明，基于Elman神经网络的商业银行信贷风险识别及评估模型能够很好地反映信贷过程中的非线性因素，准确地预测出完整的信贷风险评估指标和信用等级之间的映射关系，能快速评估和有效减低商业的信贷风险。一组实例结果显示该评估模型的准确率接近90％。     

基于遗传算法的土钉支护结构优化设计研究

通过建立三维有限元模型,对土钉支护的变形和受力性能进行分析,得出土钉力的合理分布。在此基础上,考虑土钉支护的水平位移,采用遗传算法,对土钉支护的结构优化设计进行研究,建立了土钉支护结构的优化设计数学模型,并编制了相应的计算程序。通过算例分析,并与基于极限平衡分析的优化结果相比较,得出合理的土钉支护结构设计的参考结论。     

基于遗传算法的门限自回归模型在浅层地下水位预测中的应用

提出了建立门限自回归模型的一种简便通用的方法，用遗传算法可同时优化门限值和自回归系数．实例计算的结果说明了其可行性和有效性．该方法在各种工程时序预测中具有广泛的应用价值     

动态模糊神经网络在大坝变形预报中的应用

针对静态模糊神经网络的局限性,提出了在线动态建模的模糊神经网络方法.当新增样本进入训练集之后,根据新样本对模型的贡献大小,在已有模型的基础上进行动态修正,这样可以减少建模的计算时间.新方法实现了增加样本而矩阵阶数不增加,避免了矩阵求逆运算,理论上可以提高计算效率.实例表明动态模糊神经网络方法是可行的,可实现持久预报,具有较强的适应能力和较高的预报精度,可应用于在线实时变形预报及相关领域.     

降雨空间分布模式识别

以广东省内186个站点的年降雨量分布为基础，把模式识别的聚类分析理论方法引入降雨空间特征分类研究中，进行降雨特征空间分布模式识别，以预测无雨量站点的降雨模式，从而分析水资源的空间分布模式，以便制订有效的水资源利用方案。结果表明，广东省年降雨量在空间上的分布可划分为6类模式；同时基于BP人工神经网络算法，建立了以高程、气温、蒸发量为主要参数的可判别和预测无雨量测站区域的降雨空间分布模式所属类别的模型，经过模型的学习和验证，效果较好。     

两相电流互感器配电网零序电流构造方法

中性点非直接接地配电网大多只安装A，C两相电流互感器（TA），而现有单相接地故障选线方法基本都以零序电流为基础实现。如何在两相TA配电网中准确检测并隔离接地故障线路，成为长期困扰实际运行的技术难题。通过对暂态电流的频域和时域分析，找出B相暂态电流与A，C相暂态电流的对应关系，提出在A相和C相接地时由A相和C相暂态电流构造B相暂态电流，进而得到较精确零序电流的方法。通过大量仿真实验验证了所提出的构造方法的正确性。零序电流的成功构造为实现两相TA配电网单相接地故障选线提供了基础。