SEMIPARAMETRIC MODEL SELECTION IN LARGE SAMPLES

1 IntroductionThere are obvious reasons for the popularity of linear regression among which are illter-pretation and simplicity in computation. But this does not mean that a llnear relationship isalways sufficient. In some aPpllcations, the mean is linearly related to some variables but therelation to additiona1 variables are not easily parameterized. Partly linear models become thenatural choices in such aPplicationst the linear model is drinimally altered to allow one or a fewof the indepen…     

模型选择准则的渐近性质与选择检验

模型选择分为绝对与相对两类，本文针对实用性好但理论上不甚完善的相对选择问题，归纳了其建立准则的一般性途径，证明了它的渐近统计特性，并由此提出了“选择检验”的新概念及其表达。     

估计GM(1,1)模型中参数的线性规划方法

估计GM(1,1)模型中的参数通常采用最小二乘准则,而在模型精度检验时又常采用平均相对误差。在平均相对误差达到最小准则或最大相对误差达到最小准则时,分别给出了估计GM(1,1)模型中参数的线性规划方法,并通过实例给出了不同极小化准则下数值结果的对比。数值结果表明,采用平均相对误差达到最小准则和最大相对误差达到最小准则比通常采用的最小二乘准则更合理,效果更好。     

OUTLIER SCORE TEST IN A GROWTH CURVE MODEL

ＯＵＴＬＩＥＲＳＣＯＲＥＴＥＳＴＩＮＡＧＲＯＷＴＨＣＵＲＶＥＭＯＤＥＬＦＥＩＹｕ（ＤｅｐａｒｔｍｅｎｔｏｆＡｃｃｏｕｎｔｉｎｇ，ＹｕｎｎａｎＵｎｉｖｅｒｓｉｔｙ，Ｋｕｎｍｉｎｇ６５００９１，Ｃｈｉｎａ）Ａｂｓｔｒａｃｔ：Ｓｃｏｒｅｔｅｓｔｓｔａｔｉｓ...     

MODEL SELECTION FOR LOG-LINEAR MODELS OF CONTINGENCY TABLES

In this paper, we propose an information-theoretic-criterion-based model selection procedure for log-linear model of contingency tables under multinomial sampling, and establish the strong consistency of the method under some mild conditions. An exponential bound of miss detection probability is also obtained. The .selection procedure is modified so that it can be used in practice. Simulation shows that the modified method is valid. To avoid selecting the penalty coefficient in the information criteria, an alteruative selection procedure is given.     

通过GIC准则同时诊断数据变换与异常点

针对回归分析中异常数据和变量变换相互影响的问题,从变量选择角度,结合模型选择的广义信息准则(GIC)与构造变量方法,提出了一类数据变换与异常点的同时诊断方法。该方法同时考虑由是否存在异常点以及是否需要变换所组成的四种备选模型,在某些情况下,既可以减轻异常点对数据变换的强影响,又避免了变换数据对于异常点的掩盖效应. 文章通过模拟与实例验证了该方法的有效性,并与文献中的$F$检验法、似然距离诊断法以及$F$方法进行了比较.     

我国航空客运量需求预测模型:基于随机前沿预测模型和模型平均

针对多数研究以产品或服务的历史消费量来代替不可观测的需求量而导致的需求预测出现实质性偏差的问题，本文将包含技术无效率项的随机前沿预测模型应用于航空客运量需求的预测，从而有效解决实质性偏差的问题.同时我们在此基础上引入一种模型平均权重确定方法，即通过最小化M折交叉验证准则（CV_M）确定候选模型权重.本文证明了该方法在理论上的最优性.由于模型中技术无效项的存在，我们可以同时预测航空客运量的实际产生量和需求量，实证研究也表明，相比其他常用的预测方法，该方法在预测航空客运量中长期的实际产生量上更具优势.     

非均衡模型的可解性研究

本文研究了多市场非均衡模型（ＭＤＥＭ）的可解性问题，给出微观ＭＤＥＭ可解性的一个充要条件；推广了Ｉｔｏ［１］的充分条件；对具有Ｂｕｒｋｅｔｔ双曲线聚合［２］和Ｌａｍｂｅｒｔ指数聚合［３］两类非线性形式的宏观ＭＤＥＭ，分别给出其可解性的一个充分条件，并探讨ＭＤＥＭ中一类线性溢出效应的微观基础，Ｉｔｏ［１］所分析的双市场情形是其特例．     

A RISK-AVERSE NEWSVENDOR MODEL WITH LIMITED CAPACITY AND OUTSOURCING UNDER THE CVAR CRITERION

For seasonal products,the retailers usually launch various marketing efforts,like advertising and promotion,to promote them in a selling season.While facing large demand from the customers,one should take the capacity constraint and outsourcing into consideration.Considering the shorten life cycles of most products,in this paper we adopt the traditional newsvendor model to investigate the optimal marketing effort along with optimal order quantity.We address the risk aversion issue and characterize the influence of the sellers’ risk propensity with CVaR criterion,and we develop an effective algorithm to obtain the optimal strategy.The impact of sellers’ risk propensity on the performance of the system is illustrated via numerical examples.The innovation of this paper is threefold.First,the optimal joint strategy of the marketing effort and order quantity is investigated and an efficient algorithm to find the optimal strategy is developed.Second,the capacity constraint option and an outsourcing strategy are studied jointly for excess products.Finally,the risk propensity of the seller and its influence are investigated by using the CVaR criterion,through which we obtain some new managerial insights.     

人口模型中的分岔现象

在非线性动态系统中可能出现分岔现象,分岔理论为我们研究社会经济系统中的某些异常现象提供了一种新方法。本文研究了人口模型的分岔现象,以此作为示例,体现了用分岔理论研究社会经济系统的过程与步骤。     

基于能量判据的柔性机械臂模型降阶

应用有限元法进行柔性机械臂的建模和分析非常有效，然而随着对柔性机械臂分段数的增加，系统动力学模型阶数急剧增大，需要进行模型简化。建立了基于能量范数的能量判据，将能量判据作为柔性机械臂系统模态截断的一个标准，以能量响应信息的收敛程度或者模态能量贡献作为系统模态截断的判断标准。通过与基于均衡实现的模态截断的比较，显示基于能量判据的模态截断法更适合于考虑刚柔耦合的柔性机械臂系统。对单连杆柔性机械臂有限元模型进行降阶，仿真结果表明提出方法的有效性。     

多元质量特性的稳健设计及其实现

多元质量特性稳健设计的关键在于既要确定多个质量特性整体波动的质量,又要建立方便,实用的实现途径的特点,给出多元质量特性稳健设计的一般模型,然后利用信息论中“熵”的概念和多元统计的结果,建立多元质量特性稳健设计波动源的度量准则和便于操作的实施程序。利用内、外表直积法,通过仿真分析给出了实现多元质量特性稳健设计的实施过程,结果表明该方法是可行、有效的。     

A STUDY ON THE CHAOS MODEL OF LIQUIDITY IN STOCK MARKET

For a Stock Market, the critical problem is the maintenance of its liquidity. Market liquidity can be described in various ways, in particular, in terms of the bid/offer spread and the market depth. Model of market liquidity dynamics has been proposed in Schmidt, A.B.' literate. In our study, we improve his model. On one hand, we think that trading volume is determined by the total number of traders, as well as the relations between the numbers of buyers and sellers, while the model of Schmidt only considers the first, item. On the other hand, Schmidt assumes that the number of "newcomers" in the market is in proportion to the current number of trades. However, we all know that the continual rise or fall of the price will also attract more buyers or sellers, that, is, "newcomers", into the market, which he has not taken for granted. We also prove it, to be a chaos model through analysis of Lyapunov exponent. On the assumption that price variation can be neglected, we discuss the conditions in which chao     

安全第一准则下的动态资产组合选择

考虑连续时间金融市场的最优资产组合选择问题.在Black-Scholes金融市场设置下,利用Roy提出的安全第一(SafetyFirst)准则,导出了最优常数再调整资产组合投资策略的显式表达式.还将文中的结果与Markowitz均值-方差模型的最优资产组合投资策略进行了比较,并给出概念与结论的一些经济学解释和应用.     

基于ANN预报模型的快速学习算法

提出了一种基于预报模型的变目标函数的前向神经网络快速学习算法。首先推导出一种综合目标函数从而实现极值点附近收敛速度的提高，然后导出变目标函数的局部化递扒最小二乘算法，该方法与现有同类的算法相比，可以提高收敛速度和预报精度，适用于需快速学习的系统辨识和其他应用。系统辨识的仿真实例表明了算法的优良性能。     

BLOW-UP CRITERION OF SMOOTH SOLUTIONS TO THE MHD EQUATIONS IN BESOV SPACES

In this paper we discuss the logarithmic Sobolev inequalities in Besov spaces,and show their applications to the blow-up criterion of smooth solutions to the incompressible magneto-hydrodynamics equations.     

HHM模型的多参数Hopf分岔分析

现代科学证明，不少疾病都与生物细胞膜离子通道病变有关，表现为生物模型中相关参数的异常变化，这类疾病可归类为离子通道疾病。以肌肉中的Hodgkin-Huxley模型为研究对象，选取漏电导数gⅠ和钠离子通道反电势VNa作为分岔参数，分析其Hopf分岔并探讨分岔对HH模型的影响。采用高维方程的代数判据进行Hodgkin-Huxley模型多参数动态分岔分析，简化了分析过程。并用研究结果解释相应的生理过程，试图从生物系统动态过程异变的角度探讨生理疾病的成因。     

不同借贷利率下的投资组合选择——-基于均值和VaR的效用最大化模型

用VaR代替方差来度量风险,从而把基于均值和方差的效用函数拓展为基于均值和VaR的一般二元效用函数（关于均值递增,关于VaR递减）,进而研究含无风险资产且具有不同借贷利率时投资组合选择的效用最大化模型.利用均值-VaR模型有效边界的性质,得到了一般效用函数下最大效用存在的条件及最优解的本质特征,并给出了求解的具体方法和数值算法.最后作为结论的直接应用和说明,利用中国股票市场数据给出了一个实例分析.     

The Choice of a Forecasting Model

In this article, several criteria for forecasting model selection are proposed, and some empirical comparison results are given.     

肌肉中的HH模型钠离子通道反电势的Hopf分岔分析

以肌肉中的Hodgkin-Huxley模型为研究对象，研究病理实验中有显著变化的钠离子通道反电势参数对Hodgkin-Huxley模型的影响并分析其Hopf分岔。采用高维方程的代数判据进行Hodgkin-Huxley模型单参数动态分岔分析，简化了分析过程，并用研究结果解释相应的生理过程，试图从生物系统动态过程异变的角度探讨生理疾病的成因。