Heteroscedasticity Detection and Estimation with Quantile Difference Method

When dealing with regression analysis, heteroscedasticity is a problem that the authors have to face with. Especially if little information can be got in advance, detection of heteroscedasticity as well as estimation of statistical models could be even more difficult. To this end, this paper proposes a quantile difference method (QDM) that can effectively estimate the heteroscedastic function. This method, being completely free from the estimation of mean regression function, is simple, robust and easy to implement. Moreover, the QDM method enables the detection of heteroscedasticity without any restrictions on error terms, consequently being widely applied. What is worth mentioning is that based on the proposed approach estimators of both mean regression function and heteroscedastic function can be obtained. In the end, the authors conduct some simulations to examine the performance of the proposed methods and use a real data to make an illustration.     

基于分层贝叶斯模型的稳健参数设计

针对双响应曲面模型的参数不确定性、参数之间的层次结构以及模型的异方差问题,结合分层贝叶斯建模方法提出一种新的均值-方差双响应曲面模型,并在此基础上运用所提方法实现了产品/过程的稳健参数设计。首先,建立分层贝叶斯模型,并获得参数的后验分布;其次利用Gibbs采样获得参数估计值,在此基础上构建质量损失函数,并采用遗传算法对质量损失函数进行优化求得可控因子的最佳设计水平;最后,从模型具有同方差和异方差两种情形出发,结合具体实例分别采用普通最小二乘、加权最小二乘及分层贝叶斯建立双响应曲面模型进行了比较分析,验证了所提方法的有效性。     

Regression Function Comparison for Paired Data

In this paper, the regression function comparison for paired data is studied. The proposed test statistic is based on the weighted integral of characteristic function marked by the difference of responses. There are several merits of the proposed statistic. For instance, it takes a simple V-statistic form. No bandwidth is needed. No moment conditions are required for covariates. It can be applied to covariates of any fixed dimension. The asymptotic results are also developed. It is proven that n times the proposed test statistic converges to a finite limit under the null hypothesis and the test is consistent against any fixed alternatives. Local alternative hypotheses which converge to the null hypothesis at the rate of n-1/2 are also detected. A suitable Bootstrap algorithm is also proposed for the implementation of the proposed test statistic. Simulation studies are carried out to illustrate the merits of the proposed method. A real data example is also used to illustrate the proposed testing procedures.     

支持向量机用于性能退化的可靠性评估

为解决性能退化轨迹建模中的小样本训练问题,研究了基于统计学习理论的支持向量机回归原理,提出了基于支持向量机回归模型的产品性能退化轨迹建模、寿命预测及可靠性评估方法.给出两种性能退化轨迹的支持向量机回归模型——单一模型和加权模型.实例分析表明,所提方法有较好的预测精度.加权支持向量机回归模型可在早期实现较高精度的寿命预测,提高性能退化的可靠性评估精度,从而可缩短试验时间,节约经费开支.     

Distance-Based Regression Analysis for Measuring Associations

Distance-based regression model, as a nonparametric multivariate method, has been widely used to detect the association between variations in a distance or dissimilarity matrix for outcomes and predictor variables of interest in genetic association studies, genomic analyses, and many other research areas. Based on it, a pseudo-F statistic which partitions the variation in distance matrices is often constructed to achieve the aim. To the best of our knowledge, the statistical properties of the pseudo-F statistic has not yet been well established in the literature. To fill this gap, the authors study the asymptotic null distribution of the pseudo-F statistic and show that it is asymptotically equivalent to a mixture of chi-squared random variables. Given that the pseudo-F test statistic has unsatisfactory power when the correlations of the response variables are large, the authors propose a square-root F-type test statistic which replaces the similarity matrix with its square root. The asymptotic null distribution of the new test statistic and power of both tests are also investigated. Simulation studies are conducted to validate the asymptotic distributions of the tests and demonstrate that the proposed test has more robust power than the pseudo-F test. Both test statistics are exemplified with a gene expression dataset for a prostate cancer pathway.

     

Inference for repeated measures models under heteroscedasticity

This paper considers a widely used mixed effects model in repeated measures under heteroscedasticity. Hypotheses of the equality of the fixed effects and the simultaneous confidence intervals for all pair-wise differences are discussed.A generalized F-test has been proposed to test the equality of the fixed effects in the model,but simulation results for evaluating its performance have not been shown in the literature.Moreover,the generalized F-test cannot be used to deduce the simultaneous confidence intervals for all pair-wise differences of the fixed effects.The authors propose two new p-values to test the hypotheses of equality of the fixed effects and simultaneous confidence intervals of the differences of the effects based on the generalized pivotal quantities derived in this paper.The authors also compare the empirical performances of the proposed tests and the generalized F-test. The typeⅠerror rates and powers of these tests are evaluated using the Monte Carlo simulation.The simulation studies show that the generalized F-test does not perform well in terms of typeⅠerror rate under various sample size and parameter combinations.However,the typeⅠerror probabilities of the proposed tests are always close to the nominal value.It can also be seen that the simultaneous confidence intervals perform well.     

OUTLIER SCORE TEST IN A GROWTH CURVE MODEL

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

异方差非参数回归模型均值与方差变点的小波估计与应用

金融市场中,受突发事件的影响反映资产平均收益的均值函数和反映资产收益波动的方差函数都有可能出现变点. 本文讨论了均值和方差都存在变点的异方差非参数回归模型的变点估计问题. 给出均值函数与方差函数的局部线性估计,利用函数小波系数的特性求得均值与方差变点位置的估计值并给出其收敛速度.在模拟实验中分析变点估计值的样本特性及均值变点估计与方差变点估计的相互影响.最后通过对两组股票数据的均值变点和方差变点进行估计,说明方法的有效性.     

基于联合局域相关向量机的雷达导引头状态预测方法

针对雷达导引头机电结构组成复杂、性能指标测试数据信息利用率不足、使用传统基于数据驱动的状态预测方法精度不高的问题,借鉴相关向量机(relevance vector machine,RVM)和Dempster-Shafer (D-S)证据理论,提出了一种基于证据融合和改进局域RVM的状态预测方法。首先,对标准RVM回归模型进行改进,通过构建方差高斯核函数(variance Gauss kernel function, VGKF)来提高核函数的全局性能和泛化能力;然后通过借鉴混沌序列局域预测法中邻近点个数的选取方法,利用Hannan-Quinn (H-Q)准则对训练空间预测嵌入维数进行优化,避免了主观选取的盲目性,完成了改进局域相关向量机模型(local relevance vector machine, LRVM)的构建;最后,利用具有近似退化规律的同源装备测试数据对LRVM进行了改进,通过D-S证据理论对两种模型的预测结果进行了融合,建立了联合局域相关向量机(united local relevance vector machine, U-LRVM)模型。通过对导引头相关参数的实例预测,验证了该方法的可行性和优越性。     

Generalized F-test for high dimensional regression coefficients of partially linear models

This paper proposes a test procedure for testing the regression coefficients in high dimensional partially linear models based on the F-statistic. In the partially linear model, the authors first estimate the unknown nonlinear component by some nonparametric methods and then generalize the F-statistic to test the regression coefficients under some regular conditions. During this procedure, the estimation of the nonlinear component brings much challenge to explore the properties of generalized F-test. The authors obtain some asymptotic properties of the generalized F-test in more general cases,including the asymptotic normality and the power of this test with p/n ∈(0, 1) without normality assumption. The asymptotic result is general and by adding some constraint conditions we can obtain the similar conclusions in high dimensional linear models. Through simulation studies, the authors demonstrate good finite-sample performance of the proposed test in comparison with the theoretical results. The practical utility of our method is illustrated by a real data example.     

基于加工因素和信噪比质量损失的并行公差设计

研究了加工因素与产品成本的关系,根据实际加工的统计数据,利用Matlab回归分析得到加工因素的成本综合影响系数,并给出了加工因素的成本—公差函数。针对加工工艺规划,提出采用信噪比衡量各道工序对产品质量损失的影响。利用成本—公差函数和信噪比质量损失,建立基于加工因素和信噪比质量损失的并行公差优化模型,实现了公差的并行设计,保证公差的经济性和可加工性。最后通过工程实例验证了所提出方法。     

Statistical Diagnosis for HIV Dynamics Based on Mean Shift Outlier Model

Ordinary differential equation(ODE) are widely used for quantifying HIV viral dynamics.It is interesting but challenging to estimate ODE parameters from noisy data, especially when the data have some outliers. In this study, the authors use the Mean Shift Outlier Model(MSOM) to detect outliers in HIV model based on the two-step estimation of ODE. Approximate formula for shift parameter is derived. Furthermore, a score test statistic is constructed and its approximating distribution is established. The simulation results show that: 1) The boundary points have more impact on the parameter estimation relative to interior points. 2) The proposed procedure can detect the outliers effectively. The authors illustrate the proposed approach using an application example from an HIV clinical trial and find similar pattern to the simulation studies.     

Generalized confidence regions of fixed effects in the two-way ANOVA*

The authors discuss the unbalanced two-way ANOVA model under heteroscedasticity. By taking the generalized approach, the authors derive the generalized p-values for testing the equality of fixed effects and the generalized confidence regions for these effects. The authors also provide their frequentist properties in large-sample cases. Simulation studies show that the generalized confidence regions have good coverage probabilities.     

基于高斯噪声模型的马尔可夫网络构建算法

针对小样本集构建稀疏马尔可夫网络计算量大和求解精度不高的问题，提出一种基于高斯噪声模型的迭代噪声消减（iterative noise reduction，INR）算法。该算法首先利用回归误差的高斯特性筛选相关变量，然后通过boosting方法的自回归更新策略逐步改进学习能力，最后采用赤池信息准则（Akaike information criterion，AIC）避免出现过拟合。此外，给出了自回归更新公式，实现了可控的学习错误率并分析了计算复杂度。实验结果表明，INR能有效构建高维稀疏网络，在学习效率和精度方面具有明显优势。     

Nonparametric regression under double-sampling designs

This paper studies nonparametric estimation of the regression function with surrogate outcome data under double-sampling designs, where a proxy response is observed for the full sample and the true response is observed on a validation set. A new estimation approach is proposed for estimating the regression function. The authors first estimate the regression function with a kernel smoother based on the validation subsample, and then improve the estimation by utilizing the information on the incomplete observations from the non-validation subsample and the surrogate of response from the full sample. Asymptotic normality of the proposed estimator is derived. The effectiveness of the proposed method is demonstrated via simulations.     

Robust Estimation for Partial Functional Linear Regression Model Based on Modal Regression

This paper presents a robust estimation procedure by using modal regression for the partial functional linear regression, which combines the common linear model with the functional linear regression model. The outstanding merit of the new method is that it is robust against outliers or heavy-tail error distributions while performs no worse than the least-square-based estimation method for normal error cases. The slope function is fitted by B-spline. Under suitable conditions, the authors obtain the convergence rates and asymptotic normality of the estimators. Finally, simulation studies and a real data example are conducted to examine the finite sample performance of the proposed method. Both the simulation results and the real data analysis confirm that the newly proposed method works very well.     

Testing serial correlation in semiparametric varying coefficient partially linear errors-in-variables model

The authors propose a V _{N, p} test statistic for testing finite-order serial correlation in a semiparametric varying coefficient partially linear errors-in-variables model. The test statistic is shown to have asymptotic normal distribution under the null hypothesis of no serial correlation. Some Monte Carlo experiments are conducted to examine the finite sample performance of the proposed V _{N, p} test statistic. Simulation results confirm that the proposed test performs satisfactorily in estimated size and power. This research is supported by the National Natural Science Foundation of China under Grant Nos. 10871217 and 40574003; the Science and Technology Project of Chongqing Education Committee under Grant No. KJ080609; the Doctor's Start-up Research Fund under Grant No. 08-52204; and the Youth Science Research Fund of Chongqing Technology and Business University under Grant No. 0852008.     

Test on Stochastic Block Model: Local Smoothing and Extreme Value Theory

In this paper, to obtain a consistent estimator of the number of communities, the authors present a new sequential testing procedure, based on the locally smoothed adjacency matrix and the extreme value theory. Under the null hypothesis, the test statistic converges to the type I extreme value distribution, and otherwise, it explodes fast and the divergence rate could even reach n in the strong signal case where n is the size of the network, guaranteeing high detection power. This method is simple to use and serves as an alternative approach to the novel one in Lei (2016) using random matrix theory. To detect the change of the community structure, the authors also propose a two-sample test for the stochastic block model with two observed adjacency matrices. Simulation studies justify the theory. The authors apply the proposed method to the political blog data set and find reasonable group structures.

     

基于小波的非参数回归模型均值变点的Bootstrap监测

本文首次考虑非参数回归模型均值函数结构变点的在线监测问题. 首先对回归函数的局部线性估计值进行小波变换， 基于得到的小波系数构造监测统计量， 并在原假设和备择假设下推导出监测统计量的渐近分布； 为提高监测效果， 进一步构造了Bootstrap在线监测方法， 定义了停时； 模拟结果和实例分析显示该方法可以很好地监测到变点， 并具有较短的检测延迟.     

Large Dynamic Covariance Matrix Estimation with an Application to Portfolio Allocation: A Semiparametric Reproducing Kernel Hilbert Space Approach

The estimation of high dimensional covariance matrices is an interesting and important research topic for many empirical time series problems such as asset allocation. To solve this dimension dilemma, a factor structure has often been taken into account. This paper proposes a dynamic factor structure whose factor loadings are generated in reproducing kernel Hilbert space (RKHS), to capture the dynamic feature of the covariance matrix. A simulation study is carried out to demonstrate its performance. Four different conditional variance models are considered for checking the robustness of our method and solving the conditional heteroscedasticity in the empirical study. By exploring the performance among eight introduced model candidates and the market baseline, the empirical study from 2001 to 2017 shows that portfolio allocation based on this dynamic factor structure can significantly reduce the variance, i.e., the risk, of portfolio and thus outperform the market baseline and the ones based on the traditional factor model.