基于非线性主成分分析的信用评估模型研究

传统的主成分分析（PCA）本质上是一种线性映射算法,无法有效处理非线性关系的数据。本文在分析自联想神经网络（AANN）的基础上,借鉴传统PCA方法中的序数主成分概念,提出了基于顺序自联想神经网络（SAANN）的非线性主成分分析法（NLPCA）。进一步,结合神经网络（NN）和Logisitic模型,以我国上市公司为研究对象,分别构建了基于NLPCA-NN和NLPCA-Logisitic的信用评估模型。实证结果及ROC曲线分析表明,本文构建的NLPCA相比传统的线性PCA方法能有效地实现数据的非线性特征提取与降维,提高模型预测性能。此外,实证结果还表明,在相同PCA方法处理数据的条件下,神经网络模型的信用评估效果要好于Logisitic模型。     

公司信用风险评估新模型:基于Isomap的SVM模型

本文针对样本数据较少的特点,将基于小样本的支持向量机(SVM)方法用于我国上市公司信用风险评价中。由于考虑到财务数据特征的非线性和高维性,本文采用等距特征映射(Isomap)算法对财务指标进行特征提取,以减少数据的冗余,再针对人为选择SVM参数的盲目性,应用遗传算法优化其参数。最后通过以我国上市公司财务数据为基础的实证分析表明:基于Isomap的SVM模型比BP神经网络、PCA-SVM模型具有更强的信用风险评估能力,小样本评估准确率达到91%。     

基于粗糙集和支持向量机的商业银行信用风险评估模型

结合粗糙集理论的属性约简和支持向量机(SVM)的分类机理,提出一种数据分类的混合算法;建立了基于此算法的商业银行信用风险评估模型。模型以粗糙集属性约简作为预处理器,删除冗余属性和冲突对象,但不损失有效信息;然后基于SVM进行分类建模和预测。实证表明,创建的模型分类性能良好,降低SVM分类过程的复杂度,一定程度上避免了训练模型的过拟合现象。通过与SVM和神经网络模型的比较,证实该方法用于信用风险评估的有效性。     

贝叶斯非线性分层模型在多元索赔准备金评估中的应用

本文将贝叶斯非线性分层模型应用于基于不同业务线的多元索赔准备金评估中,设计了一种合适的模型结构,将非线性分层模型与贝叶斯方法结合起来,应用WinBUGS软件对精算实务中经典流量三角形数据进行建模分析,并使用MCMC方法得到了索赔准备金完整的预测分布。这种方法扩展并超越了已有多元评估方法中最佳估计和预测均方误差估计的研究范畴。在贝叶斯框架下结合后验分布实施推断对非寿险公司偿付能力监管和行业决策具有重要作用。     

门限分位数自回归模型的预测方法及应用

金融经济系统预测是宏观经济管理的重要问题,系统中大多数变量具有非线性与异质性等特征,门限分位数自回归（TQAR）模型能够较好地揭示这一特征。本文研究TQAR模型的预测技术,给出其条件分位数预测和条件密度预测方法。数值模拟结果表明,与传统的门限均值自回归模型（TAR）和分位数自回归（QAR）模型相比,TQAR模型在预测的精度和准度方面更具优势。文章使用TQAR模型研究中国通货膨胀的非线性动态特征,并在此基础上预测通货膨胀的波动趋势。实证结果表明,TQAR模型不仅能够揭示通货膨胀的门限效应和异质效应,提供比TAR和QAR模型更高的预测精准度,而且能够通过条件密度预测曲线,细致刻画通货膨胀条件分布的位置、散布与形状等全景信息,从而为宏观经济政策的制定和调整提供科学合理的决策依据。     

基于残差函数主成分的相依函数型回归模型估计及金融应用

本文研究具有相依特征的函数型数据的函数表示方法及其回归模型的估计方法。首先基于分整理论提出基于残差函数主成分的函数表示方法,然后利用残差函数主成分对回归系数函数正则化表示,最后把函数曲线和回归系数函数代入相依函数型回归模型进行估计,并通过蒙特卡洛模拟和金融实例分析来评估参数估计和样本外预测的准确性。研究发现,相比现有的基于协方差函数和长期协方差函数的函数主成分估计方法,本文提出的基于残差函数主成分的估计方法具有良好的有限样本性质,回归系数函数估计更准确、样本外预测效果更好;基于高频数据的股市开盘价预测实证研究表明本文方法的样本外预测精度最高。与现有方法相比,本文提出的基于残差函数主成分的估计方法,既考虑函数型数据的相依特征,又避免长期协方差函数估计时面临的核函数和窗宽选择问题,为经济金融等领域具有相依特征的函数型数据提供一种函数表示方法,丰富函数型回归模型理论,为其他函数型回归模型的拓展提供借鉴。     

我国通货膨胀福利成本的再估计

本文在货币效用模型（MIU）框架下研究我国通货膨胀福利成本问题,理性经济人的动态最优化选择行为产生了一组随机非线性欧拉方程。利用这些方程,在稳定状态下推导出估计通货膨胀福利成本的方法。实证分析以欧拉方程为基础,利用广义矩方法（GMM）估计模型参数,并进而估计出我国1995-2012年的通货膨胀福利成本。我们发现名义利率比通货膨胀率对通货膨胀福利成本的影响更大,在低通货膨胀水平,通货膨胀率对福利成本影响更大。本文阐述的估计通货膨胀福利成本方法优势在于能够同时考虑通货膨胀率和名义利率两个变量的影响,而传统估计方法只能利用名义利率作为持有货币的机会成本变量来估计通货膨胀福利成本。     

有序选择因子结构模型的MCMC估计及应用

为了评估中国分布类政策效应,根据中国微观数据的变量可得性,本文在Heckman等构建的因子结构模型基础上,将Heckman基准模型中的连续型测度方程调整为离散型有序选择模型,建立了有序选择因子结构模型,并推导出MCMC估计方法。运用该方法,结合中国样本数据,本文对高等教育的分布类政策效应进行了实证估计。有序选择因子结构模型及其MCMC估计方法对于经济政策的分布类效应评估具有普遍的理论适应性和实际应用价值。     

基于LS-SVM的小微企业信用评估研究

构建并优化小微企业信用评估技术已经成为商业银行开展小微业务必然选择。基于小微企业内在特征,设计以小微企业现金流信息为违约触发机制的小微企业信用评估指标体系,构建最小二乘支持向量机模型（LS SVM）,运用某国有控股银行的小微企业贷款微观数据证实该模型能够相对提高预测精确度和稳定性。应用LS SVM构建的小微企业信用评估指标体系以及评估模型有助于提升银行对小微企业的了解程度和小微业务风险管理能力,减轻信息不对称,在一定程度能够化解供给型信贷配给导致的小微企业融资难问题。     

基于BP网络的银行信用风险评估数据的多视角填补方法

信用风险评估作为降低借贷双方信息不对等的手段,在当代金融领域十分重要。然而用于银行信用风险评估的原始数据往往会有缺失或很难获取,降低了评估模型的有效性。如何精确获取有效数据和去除不可用数据成为提高评估模型效果的一个重要且极具挑战力的环节。目前相关方面的研究多是从单视角预测缺失值,未考虑各个指标之间的关系。因此,文章提出了一种基于BP(Back Propagation)神经网络的多视角填补方法,充分利用各个指标之间的关联,针对含缺失值的指标建立多个分类器,并利用BP神经网络强大的非线性映射能力实现预测。实际数据的实验表明,相较于常用的银行信用风险评估数据填补方法,本文提出的填补方法有着较佳的效果。     

Development in Information and Communications Technology (ICT) Statistics at the Australian Bureau of Statistics

This paper outlines the ICT statistical developments that the ABS has undertaken over recent years. These developments have been on both the supply side (ICT industry) and the demand side (use of IT by sector). The content of ABS ICT surveys has changed significantly over the last few years and are expected to develop further in line with industry changes and emerging policy needs.
Recent developments in household collections have been the inclusion of IT use questions in the 2001 Census of Population and Housing and in various ABS social surveys such as Children's Participation in Culture and Leisure Activities. Annual business IT use surveys are now conducted and include topics such as Internet commerce and IT security. Biennial ICT industry surveys are conducted to provide data on ICT production, imports and exports, ICT industry employment, performance and structure information. An Internet activity survey is run every six months and provides regional Internet access details and infrastructure details on Internet service providers.
A major development for 2002–03 will be the compilation of an ICT satellite account. ABS intends to produce an ICT Information Development Plan to guide its future statistical development work in the ICT field.     

Value at Risk in the Suburbs: Eminent Domain and the Geographical Politics of the US Foreclosure Crisis

Several US counties and local governments have recently considered a novel solution to the foreclosure crisis. They plan to use eminent domain to compel the owners of mortgage debt—and specifically of private‐label mortgage‐backed securities—to sell the debt to the government at a price reflecting the loan's market value. The government would then restructure the debt and resell it to new investors. The plans are striking because—in contrast to both development‐driven eminent domain and the federal subprime bank bailout—they would force investors to assume asset devaluation and increased long‐term risk. Notably, the plans have emerged as an instance of financialization‐focused politics in suburbs and suburban cities of color, specifically majority‐black and ‐Latino/a suburbs. Local support for the plans, we argue, is rooted in the long‐term disinvestment of these ‘suburbs of exception', which became targets of subprime lending and eventually sites where the ‘financial exception' has been localized. But these demographic shifts, fragmentation and fiscal pressures have at the same time created a suburban political terrain in which the plans have gained their strongest political support.     

Towards a Theory of Serendipity: A Systematic Review and Conceptualization

Serendipity – the notion of making surprising and valuable discoveries – plays a major role in the success of individuals and organizations alike. Previous research has established the importance of serendipity and identified important individual- and organizational-level antecedents. However, the literature has been dispersed and the boundaries of the concept have been blurry, leading to a lack of conceptual clarity and structure, and thus limiting validity and managerial actionability. Based on a systematic literature review, I synthesize existing management-related research on serendipity and explicate the emergence and composition of serendipity in the organizational context. I first identify three necessary conditions that differentiate serendipity from related concepts such as luck or targeted innovation: agency, surprise, and value. Then, I draw from the literature on sensemaking, event-based theorizing, and quantum-based approaches to management to conceptualize the process of cultivating serendipity in the organizational context as a process of enabling potentiality and materialization, and develop a multi-level theory of (cultivating) serendipity. This conceptualization contributes to our collective understanding of how, why, and when (i.e., under what conditions) organizations can leverage the value in the unexpected, which opens up fruitful avenues for further research.     

Accountability of NGOs in Bangladesh A critical overview

NGOs in Bangladesh have been recognized as effective change agents in the socio-economic arena throughout the world. Their contributions in micro credit, non-formal education and primary health care are widely known. NGOs play a significant role in society. NGO accountability, as a result, has become a critical issue. NGO accountability in Bangladesh has been examined here in relation to patrons, clients and themselves. Of the three relationships only accountability to patrons is in a near satisfactory state. Accountability to clients is handicapped by lack of opportunities of beneficiaries to participate meaningfully in NGO policy making. Accountability to themselves is adversely affected due to the adherence to a strictly hierarchical system and the influence of founding and charismatic NGO leaders.     

Towards an Asian model of human resource management? A comparative analysis of China,Japan and South Korea

There has been much discussion regarding the possible decline and weakening of national systems of human resource management (HRM). Yet, culture and institutions are often cited as the major stumbling blocks to this outcome. Such constraints may dissipate in the analysis, particularly if ‘regional clusters’ are considered where geographic and cultural closeness and ‘openness’ to similar economic pressures exist. In such cases, it might be postulated that convergence in HRM will occur. We argue that China, Japan and South Korea represent such a cluster. This article examines these countries, to see if a degree of convergence is taking place and if it is towards an identifiable ‘Asian’ model of HRM. A model of change is presented that distinguishes between levels of occurrence and acceptance. Details of the development and practice of HRM in each country are then set out. The article ends with a discussion and implications section and a brief conclusion.     

The Creation of the Office for National Statistics

A new Office for National Statistics was created in the United Kingdom in 1996 following a series of developments designed to strengthen the relevance and integrity of statistics. This paper describes these fundamental changes and sets out a blueprint for future progress. It also sets current developments in the United Kingdom in an historical and international context.     

What helps? Women engineers' accounts of staying on

We have considerable understanding of the obstacles that women engineers encounter and the reasons that they leave the field, but we know less about what enables them to remain. Adopting an interpretivist approach, this article examines how a group of British women engineers in two FTSE 100 companies account for “staying on” in their male‐dominated work settings. We delineate four specific forms of help that facilitate women's retention in the field. We argue that exposure to help leads to women developing a habitus that enables them to continue working in engineering. To conclude, we draw on our findings to outline HR practices that will facilitate supportive relationships in the workplace and pave the way towards developing more positive organisational climates.     

论现代物流企业的发展观

物流业经过诸多年的发展目前已经成为最热门的行业之一,物流企业的注册数量也与日俱增,达到前所未有的规模。毋庸置疑,任何企业都希望自己迅速做强做大,并成为真正现代物流企业的典型代表,要实现这一目标一个首要问题就是要解决现代物流企业的发展观问题。     

浅谈大体积混凝土裂缝的控制

大体积混凝土的裂缝问题是实际工程中长期困扰工程技术人员的问题,其控制技术的研究是混凝土结构研究的热点问题,具有重大的学术价值和潜伏的工程背景。     

Supply Chain Strategies in the UK Fashion Industry—The Rhetoric of Partnership and Realities of Power

This paper focuses on the fashion industry, one characterised by issues such as dramatic shifts in the scale and power of major retail buyers in the market, the advent of retailer own brands, and the nature of sourcing and supply chain decisions, issues that are increasingly global in nature. The aim of this research is to explore the nature of relationships between UK high street multiple fashion retailers and their contracted suppliers, many of whom are entrepreneurial firms by most definitions of the term. Four core themes emerge from the literature and provide a framework for the research, namely, power, process, partnership, and people. The research approach was qualitative, and conducted over a period of twelve months. The paper ends with an agenda for future research.