Deep learning for modeling the collection rate for third-party buyers

This study evaluates a wide range of machine learning techniques such as deep learning, boosting, and support vector regression to predict the collection rate of more than 65,000 defaulted consumer credits from the telecommunications sector that were bought by a German third-party company. Weighted performance measures were defined based on the value of exposure at default for comparing collection rate models. The approach proposed in this paper is useful for a third-party company in managing the risk of a portfolio of defaulted credit that it purchases. The main finding is that one of the machine learning models we investigate, the deep learning model, performs significantly better out-of-sample than all other methods that can be used by an acquirer of defaulted credits based on weighted-performance measures. By using unweighted performance measures, deep learning and boosting perform similarly. Moreover, we find that using a training set with a larger proportion of the dataset does not improve prediction accuracy significantly when deep learning is used. The general conclusion is that deep learning is a potentially performance-enhancing tool for credit risk management.     

Non-linear Gaussian sovereign CDS pricing models

Prior literature indicates that quadratic models and the Black–Karasinski model are very promising for CDS pricing. This paper extends these models and the Black [J. Finance 1995, 50, 1371–1376] model for pricing sovereign CDS’s. For all 10 sovereigns in the sample quadratic models best fit CDS spreads in-sample, and a four factor quadratic model can account for the joint effects on CDS spreads of default risk, default loss risk and liquidity risk with no restriction to factors correlation. Liquidity risk appears to affect sovereign CDS spreads. However, quadratic models tend to over-fit some CDS maturities at the expense of other maturities, while the BK model is particularly immune from this tendency. The Black model seems preferable because its out-of-sample performance in the time series dimension is the best.     

高校跨学科创新团队知识互动程度评价研究

知识互动程度评价对改善高校跨学科创新团队知识互动具有重要意义。在界定高校跨学科创新团队知识互动及知识互动程度内涵的基础上，创新性地从知识互动深度、互动广度、互动时间3个维度建立知识互动程度评价指标体系。进一步地，采用三角模糊权重、熵权和模糊积分相结合的方法，构建知识互动程度评价模型。最后，选取一所高校跨学科创新团队进行实例分析，验证评价模型的可操作性，可为高校跨学科创新团队知识互动程度评价提供一个新视角。     

数学素质教育的发现法

素质教育倡导数学教学应重视学生的主观能动性，力求使学生亲身经历数学问题的发现过程，体验研究成果的喜悦。文章就发现法教学从理论到实践阐述了这种教学方式的五种途径。     

Cardinality-based equality of opportunities

We study possible rankings of opportunity profiles. An opportunity profile is a list of opportunity sets, one for each agent in the society. We compare profiles on the basis of the notion of “equality of opportunities”. Our main results show the necessary and sufficient conditions for this comparison to be made using exclusively the information provided by two cardinal measures: the number of common alternatives in all sets of a given profile and the difference between the number of alternatives in the individual sets. We also show that, under additional axioms, we can combine these criteria only in a weighted procedure or in some lexicographic refinements of a weighted procedure.      

A comparative site evaluation study for relocating Japanese government agencies out of Tokyo

As a national project of Japan, plans for the relocation of several government agencies out of Tokyo have been ongoing. This paper is concerned with the problem of site selection for this project. The National Land Agency, the agency responsible for this project, has declared that the process of site selection should be rational, open to the public and easily understandable. In an effort to meet these requirements, we propose a consensus-making method for reaching a group decision, based on a combination of the analytic hierarchy process (AHP) and the Assurance Region model of data envelopment analysis (DEA). Several strategic uses of these methodologies, e.g., Delphi procedures, are also discussed. Based on these analyses, the “Wise Men” Committee for deciding the best site has chosen two from among ten contenders, one from the North-East and the other from the Central part of Japan, as candidates for the best relocation site. We could not discriminate between the two with respect to the multiple criteria chosen for evaluating sites. The Committee recommended the two sites to the Prime Minister at the end of December 1999. The Prime Minister reported this conclusion to the Diet. This topic is currently the focus of political discussions in the Diet. In this report, the authors summarize the decision-making processes that the Wise Men Committee followed, putting emphasis on the methodological aspects.     

基于AHP＋熵组合的相对发达地区新乡镇建设评价——以苏南三镇为例

面对新农村建设活跃和城市化进程加速，新乡镇建设也初见端倪，选取苏南三镇作为实证地点，构建相对发达地区新乡镇建设评价体系，利用主观赋权的层次分析法（AHP）和客观赋权的熵权法进行组合评价。     

A generalization of Condorcet’s Jury Theorem to weighted voting games with many small voters

We extend Condorcet’s Jury Theorem (Essai sur l’application de l’analyse à la probabilité des décisions rendues à la pluralité des voix. De l’imprimerie royale, 1785) to weighted voting games with voters of two kinds: a fixed (possibly empty) set of ‘major’ voters with fixed weights, and an ever-increasing number of ‘minor’ voters, whose total weight is also fixed, but where each individual’s weight becomes negligible. As our main result, we obtain the limiting probability that the jury will arrive at the correct decision as a function of the competence of the few major players. As in Condorcet’s result the quota q = 1/2 is found to play a prominent role. I wish to thank Maurice Koster, Moshé Machover, Guillermo Owen and two anonymous referees for helpful comments.     

Turbo码解码过程中的加权算法

本文通过调整迭代解码过程中系统位接收值的加权系数，提出了一种Turbo码加权迭代解码算法。该算法改变了迭代运算后Turbo码解码器输出软值中系统位接收值信息和它的外部估计信息的比重，使Turbo码无论在低信噪比或是在高信噪比时均具有优良的纠错性能。仿真结果显示，采用Turbo码加权迭代解码算法，不仅能提高Turbo码的收敛速度，而且能进一步降低Turbo码解码时的地板值，使Turbo码的比特误码率在高、低信噪比时都能够得到进一步改善。     

低负荷维氏硬度法的负荷选择

本文按照维氏硬度测定的技术条件,将基本公式H_v=1854P/D~2加以变换,从而得到一些较为简单的数学关系式.利用这些关系式,可以对不同金属薄带材料,很容易计算出较合理的负荷大小,既能保证不被"打穿",也使视场内有一个适合观测的压痕图象.