时间序列建模中滞后阶数选取准则函数的检测效力及其特征

研究了时间序列建模中滞后阶数选取准则函数的检测效力及其特征.采用Monte-Carlo模拟对于所有的1阶和2阶自回归模型进行了系统仿真,研究结论表明:1)随着样本容量的逐渐增大,SIC准则函数正确判断出平稳序列滞后阶数的概率将显著增大,并渐近依概率收敛到1,而AIC准则函数正确判断出平稳序列滞后阶数的概率将逐渐增大,但并不能依概率收敛到真值.2)在其它外界条件不变情况下,所有准则函数正确判断概率与序列的平稳性无显著变化关系.3)样本容量越小,AICC检测效力越显著高于SIC、AIC检测效力;随着样本容量的逐渐增大,SIC检测效力逐渐高于AICC、AIC检测效力,但此时AICC、AIC检测效力相当.     

AUTOMATIC SELECTION OF PARAMETERS IN SPLINE REGRESSION VIA KULLBACK－LEIBLER INFORMATION

ＡＵＴＯＭＡＴＩＣＳＥＬＥＣＴＩＯＮＯＦＰＡＲＡＭＥＴＥＲＳＩＮＳＰＬＩＮＥＲＥＧＲＥＳＳＩＯＮＶＩＡＫＵＬＬＢＡＣＫ－ＬＥＩＢＬＥＲＩＮＦＯＲＭＡＴＩＯＮＡＵＴＯＭＡＴＩＣＳＥＬＥＣＴＩＯＮＯＦＰＡＲＡＭＥＴＥＲＳＩＮＳＰＬＩＮＥＲＥＧＲＥＳＳＩＯ...     

风险度量半参数变系数复合Expectile回归模型及应用

本文结合半参数变系数回归模型、期望分位数风险价值(EVaR)的思想以及充分利用多个Expectile信息能提高参数估计效率的假设,提出了一类半参数变系数复合Expectile回归模型,并对该模型进行了估计,建立了所提出复合Expectile回归(CER)估计的大样本性质.针对该模型既含有参数部分也含有非参数部分的特征,采用了方便计算的三步估计方法.通过数值模拟也发现,当误差为厚尾或非对称分布时,在均方根误差(RMSE)的标准下,所提出的CER估计大大优于最小二乘(LS)估计和简单的Expectile回归(ER)估计.另外,本文还应用所发展的理论分析了我国货币政策对上证综指的影响.     

GMDH-Based Outlier Detection Model in Classi?cation Problems

In many practical classification problems, datasets would have a portion of outliers, which could greatly affect the performance of the constructed models. In order to address this issue, we apply the group method of data handing(GMDH) neural network in outlier detection. This study builds a GMDH-based outlier detection(GOD) model. This model first implements feature selection in the training set L using GMDH neural network. Then a new training set L can be obtained by mapping the selected key feature subset. Next, a linear regression model can be constructed in the set L by ordinary least squares estimation. Further, it eliminates a sample from the set L randomly every time, and then rebuilds a linear regression model. Finally, outlier detection is realized by calculating Cook's distance for each sample. Four different customer classification datasets are used to conduct experiments. Results show that GOD model can effectively eliminate outliers, and compared with the five existing outlier detection models, it generally performs significantly better. This indicates that eliminating outliers can effectively enhance classification accuracy of the trained classification model.     

Heckman-Tobit模型的半参数估计

Heckman-Tobit模型可以同时处理样本选择问题和删失数据问题,是一个重要的微观计量模型.本文根据结果变量的条件生存函数所满足的性质,提出Heckman-Tobit模型的一种半参数估计方法.这种方法通过积分的形式,有效地利用了结果变量整个条件分布的信息.在一些正则性条件下,本文证明了所提出的半参数估计量的相合性和渐近正态性.其渐近性质的成立不依赖于扰动项的具体分布.数值模拟实验的结果表明,本文的半参数估计量具有优越的有限样本性质,且当扰动项服从非正态分布时优于最大似然估计量.     

热线式空气质量流量传感器Hammerstein模型结构辨识

利用稳态和动态校准信息,辨识了基于Hammerstein模型的热线式空气质量流量(MAF)传感器模型结构.模型辨识采用两步法,在用多项式逼近静态非线性特性的基础上,动态线性环节分别选取ARX模型、输出误差(OE)模型和Box-Jenkins(BJ)模型,采用交叉准则法进行参数估计和阶次选择,模型的残差分析和用验证数据对模型交叉检验的结果表明,最终输出误差(FOE)准则和最终预报误差(FPE)准则选择的阶次一致,基于预测误差法的3阶OE和BJ模型均可用于热线式MAF传感器Hammerstein模型动态线性环节的建模.     

Variable Selection via Generalized SELO-Penalized Cox Regression Models

The seamless-L_0(SELO) penalty is a smooth function that very closely resembles the L_0 penalty, which has been demonstrated theoretically and practically to be effective in nonconvex penalization for variable selection. In this paper, the authors first generalize the SELO penalty to a class of penalties retaining good features of SELO, and then develop variable selection and parameter estimation in Cox models using the proposed generalized SELO(GSELO) penalized log partial likelihood(PPL) approach. The authors show that the GSELO-PPL procedure possesses the oracle property with a diverging number of predictors under certain mild, interpretable regularity conditions. The entire path of GSELO-PPL estimates can be efficiently computed through a smoothing quasi-Newton(SQN) with continuation algorithm. The authors propose a consistent modified BIC(MBIC) tuning parameter selector for GSELO-PPL, and show that under some regularity conditions, the GSELOPPL-MBIC procedure consistently identifies the true model. Simulation studies and real data analysis are conducted to evaluate the finite sample performance of the proposed method.     

Profile Statistical Inference for Partially Linear Additive Models with a Diverging Number of Parameters

This paper considers partially linear additive models with the number of parameters diverging when some linear constraints on the parametric part are available. This paper proposes a constrained profile least-squares estimation for the parametric components with the nonparametric functions being estimated by basis function approximations. The consistency and asymptotic normality of the restricted estimator are given under some certain conditions. The authors construct a profile likelihood ratio test statistic to test the validity of the linear constraints on the parametric components,and demonstrate that it follows asymptotically chi-squared distribution under the null and alternative hypotheses. The finite sample performance of the proposed method is illustrated by simulation studies and a data analysis.     

Adjusted Empirical Likelihood Estimation of Distribution Function and Quantile with Nonignorable Missing Data

This paper considers the estimation problem of distribution functions and quantiles with nonignorable missing response data. Three approaches are developed to estimate distribution functions and quantiles, i.e., the Horvtiz-Thompson-type method, regression imputation method and augmented inverse probability weighted approach. The propensity score is specified by a semiparametric exponential tilting model. To estimate the tilting parameter in the propensity score, the authors propose an adjusted empirical likelihood method to deal with the over-identified system. Under some regular conditions, the authors investigate the asymptotic properties of the proposed three estimators for distribution functions and quantiles, and find that these estimators have the same asymptotic variance. The jackknife method is employed to consistently estimate the asymptotic variances. Simulation studies are conducted to investigate the finite sample performance of the proposed methodologies.     

A Constrained Interval-Valued Linear Regression Model: A New Heteroscedasticity Estimation Method

Linear regression models for interval-valued data have been widely studied. Most literatures are to split an interval into two real numbers, i.e., the left- and right-endpoints or the center and radius of this interval, and fit two separate real-valued or two dimension linear regression models. This paper is focused on the bias-corrected and heteroscedasticity-adjusted modeling by imposing order constraint to the endpoints of the response interval and weighted linear least squares with estimated covariance matrix, based on a generalized linear model for interval-valued data. A three step estimation method is proposed. Theoretical conclusions and numerical evaluations show that the proposed estimator has higher efficiency than previous estimators.

     

Semiparametric Bayesian inference for accelerated failure time models with errors-in-covariates and doubly censored data

This paper proposes a Bayesian semiparametric accelerated failure time model for doubly censored data with errors-in-covariates. The authors model the distributions of the unobserved covariates and the regression errors via the Dirichlet processes. Moreover, the authors extend the Bayesian Lasso approach to our semiparametric model for variable selection. The authors develop the Markov chain Monte Carlo strategies for posterior calculation. Simulation studies are conducted to show the performance of the proposed method. The authors also demonstrate the implementation of the method using analysis of PBC data and ACTG 175 data.     

Sequential Feature Screening for Generalized Linear Models with Sparse Ultra-High Dimensional Data

This paper considers the iterative sequential lasso(ISLasso) variable selection for generalized linear model with ultrahigh dimensional feature space. The ISLasso selects features by estimated parameter sequentially iteratively for the second order approximation of likelihood function where the features selected depend on regulatory parameters. The procedure stops when extended BIC(EBIC)reaches a minimum. Simulation study demonstrates that the new method is a desirable approach over other methods.     

船舶姿态运动实时预报算法研究

根据大型舰船在随机海浪作用下的非平稳运动特性,提出基于AR模型的实时建模预报方法,详细讨论了改进隅角实时快速定阶算法和RLS递推在线参数辨识方法。针对典型航行工况,对船舶纵摇运动进行了仿真研究,并与AIC定阶算法进行了比较。仿真结果表明:AR算法适用于舰船在非平稳运动情况下的建模预报。AIC准则、改进隅角两种定阶方法均可实现AR模型阶数的在线估计问题,预报长度均可达到7-10秒,但改进隅角定阶算法简单,独立性好,易于编程实现,预报实时性好。该方法在理论和工程应用方面具有重要的意义。     

动态空间自回归固定效应面板模型的估计与应用

文章首次探寻了在空间系统稳定以及n和T均为很大的情况下,DSAC固定效应面板模型的拟极大似然估计量的渐近性质.研究发现:运用转换法估计时,在一般情况下得到拟极大似然估计量存在O(1/T)阶的偏差,当(n-1)/T→0时,转换法得到的估计量以√(n-1)/T的速度一致地收敛于真值,当(n-1)/T→∞时,估计量以T的速度收敛至一个退化分布;用直接法估计时,在一般情况下得到的估计量会产生max(O(1/T),O(1/n))阶的偏差,当n/T→0和n/T→∞时,估计量分别以n和T的速度收敛至不同的退化分布;偏差修正估计量比拟极大似然估计量具有更好的有限样本性质:当n/T³→0时,转换法得到的偏差修正估计量以√(n-1)/T1的速度一致地收敛于真值,当n/T³和n³/T同时趋于0时,直接法得到的偏差修正估计量以√nT的速度一致地收敛于真值;直接法可以一致地估计个体效应和时间效应而转换法不能;当扰动项存在空间相关结构时DSAC固定效应面板模型的有限样本性质优于DSAR面板模型;最后用一个实证研究的例子表明了DSAC模型的应用价值.     

Unscented卡尔曼滤波-卡尔曼滤波算法

针对条件线性高斯状态空间模型,提出unscented卡尔曼滤波-卡尔曼滤波unscented Kalman filte-ring-Kalman filtering,UKF-KF算法,该方法用UKF估计条件线性高斯状态空间模型中的非线性状态,用KF估计线性状态。为了有效地融合UKF和KF估计的后验状态分布,将蒙特卡罗方法应用于KF估计的线性状态均值和方差,获得了与UKF sigma点相同数量的后验线性状态估计分布的样本,然后将这些样本与UKF中sigma点进行合成去获得系统中非线性状态的估计。该算法应用于机动目标跟踪的仿真结果表明:与Rao-Blackwellized粒子滤波器(Rao-Blackwellized particle filter,RBPF)相比,该算法虽在估计精度上略有下降,然而计算时间明显降低,有效提高了实时性。     

Electricity Consumption and Its Impact Factors: Based on the Nonparametric Model

The security of power consumption demand is an essential issue in maintaining a rapid economic development in China. Owing to the advantages such as the robustness and favorable estimation performances, the semiparametric model and nonparametric models have obtained wide popularity. In this article, based on the semiparametric and nonparametric models, a new estimation model is created, and the linear and nonlinear effects of influencing the power consumption are investigated. The results indicate that the economic growth, population, and economic structure play a vital role in the power consumption; the impact of the power price index on the electricity consumption is small, which can-not offset rapid growth of the electricity consumption that the economic growth, population and economic structure bring; the energy utilization efficiency still stays at a low level in China.     

陀螺随机漂移时间序列建模方法研究

提出了残差自校正的RELS算法，在ARMA模型参数估计的过程中引入虚拟白噪声，来补偿参数估计过程中的由于噪声未知而引入的误差，并将DDS法与AIC准则相结合确定模型阶数。仿真结果表明所确定的模型能够精确地反映陀螺漂移趋势。     

模糊软集合理论在税收组合预测中的应用

结合模糊软集合理论建立税收收入的组合预测模型,根据税收收入的特点,代表性地选择了Elman回归神经网络模型、含政策虚拟变量的自回归模型、ARIMA(1,1,1)的时间序列模型、多因素SVM回归模型这四种模型作为组合预测中的单一模型,并以1980年到2008年的税收收入等相关数据为背景进行了说明和分析.结果表明该组合预测模型能有效减小预测误差,为税收工作实践提供了一个应用研究工具,并推广和丰富了软集合理论在税收经济模型研制中的实际应用.     

杂志订阅的0-1目标规划模型及其应用

论述了报刊杂志订阅选择或取消订阅问题的重要性 ,对常用的杂志订阅选择方法进行了综合分析 .为克服常用方法的缺陷即无力解决具有资源约束的、多目标相互冲突且计量单位不可比的杂志选择问题 ,本文提出了杂志订阅选择的 0 -1目标规划模型 .最后 ,本文以某大学的实际数据对模型进行了上机运行 .结果表明 ,目标规划法是一种科学合理的杂志订阅选择的较满意方法.     

A branch-and-cut approach to portfolio selection with marginal risk control in a linear conic programming framework

Marginal risk represents the risk contribution of an individual asset to the risk of the entire portfolio In this paper, we investigate the portfolio selection problem with direct marginal risk control in a linear conic programming framework. ＇The optimization model involved is a nonconvex quadratically constrained quadratic programming （QCQP） problem. We first transform the QCQP problem into a linear conic programming problem, and then approximate the problem by semidefinite programming （SDP） relaxation problems over some subrectangles. In order to improve the lower bounds obtained from the SDP relaxation problems, linear and quadratic polar cuts are introduced for designing a branch-and-cut algorithm, that may yield an e -optimal global solution （with respect to feasibility and optimality） in a finite number of iterations. By exploring the special structure of the SDP relaxation problems, an adaptive branch-and-cut rule is employed to speed up the computation. The proposed algorithm is tested and compared with a known method in the literature for portfolio selection problems with hundreds of assets and tens of marginal risk control constraints.