A moment-matching robust collaborative optimization method

Fenfen Xiong,Gaorong Sun,Ying Xiong,Shuxing Yang 대한기계학회 2014 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.28 No.4

Robust collaborative optimization (RCO) is a widely used approach to design multidisciplinary system under uncertainty. In most ofthe existing RCO frameworks, the mean of the state variable is considered as auxiliary design variable and the implicit uncertainty propagationmethod is employed for estimating their uncertainties (interval or standard deviation), which are then used to calculate uncertaintiesin the ending performances. However, as repeated calculation of the global sensitivity equations (GSE) is demanded during the optimizationprocess of the existing approaches, it is typically very cumbersome or even impossible to obtain GSE for many practical engineeringproblems due to the non-smoothness and discontinuity of the black-box-type analysis models. To address this issue, a new RCOmethod is proposed in this paper, in which the standard deviation of the state variable is introduced as auxiliary design variable in additionto the mean. Accordingly, interdisciplinary compatibility constraint on the standard deviation of state variable is added to enhancethe design compatibility between various disciplines. The effectiveness of the proposed method is demonstrated through two mathematicalexamples. The results generated by the conventional robust all-in-one (RAIO) approach are used as benchmarks for comparison. Ourstudy shows that the optimal solutions produced by the proposed RCO method are highly close to those of RAIO while exhibiting goodinterdisciplinary compatibility.

Testing for covariance matrices in time-varying coefficient panel datamodels with fixed effects

Ranran Chen,Gaorong Li,Sanying Feng 한국통계학회 2020 Journal of the Korean Statistical Society Vol.49 No.1

In this paper, we study the tests for sphericity and identity of covariance matrices in time-varying coefficient high-dimensional panel data models with fixed effects. In order to construct the effective test statistics and avoid the influence of the unknown fixed effects, we apply the difference method to eliminate the dependence of the residual sample, and further construct test statistics using the trace estimators of the covariance matrices. For the estimators of the coefficient functions, we use the local linear dummy variable method. Under some regularity conditions, we study the asymptotic property of the estimators and establish the asymptotic distributions of our proposed test statistics without specifying an explicit relationship between the cross-sectional and the time series dimensions.We further show that the test statistics are asymptotic distribution-free. Subsequently simulation studies are carried out to evaluate our proposed methods. In order to assess the performance of our proposed test method, we compare with the existing test methods in panel data linear models with fixed effects.

Statistical inference for the unbalanced two-way error component regression model with errors-in-variables

Lili Yue,Gaorong Li,Junhua Zhang 한국통계학회 2017 Journal of the Korean Statistical Society Vol.46 No.4

In this paper, we investigate the estimation and testing problems of unbalanced twoway error component regression model with errors-in-variables. The estimation of the unknown parameter is given based on the bias-corrected technique, and the asymptotic property of the resulting estimator is shown under some regularity conditions. For the hypothesis testing problem of restricted condition, a test statistic is constructed through the difference of the corrected residual sums of squares under the null and alternative hypotheses. Then we develop an adjusted test statistic, which follows an asymptotically standard Chi-squared distribution. Some simulation studies and a real data analysis are carried out to assess the performance of the proposed methods.

Weighted bias-corrected restricted statistical inference for heteroscedastic semiparametric varying-coefficient errors-in-variables model

Zhang Weiwei,Li Gaorong 한국통계학회 2021 Journal of the Korean Statistical Society Vol.50 No.4

In this paper, we consider statistical inference for a heteroscedastic semiparametric varying-coefficient partially linear model with measurement errors in the nonparametric component when exact linear restriction on the parametric component is assumed to hold. Two types of weighted bias-corrected restricted estimators of the parametric and nonparametric components are proposed based on a bias-corrected estimator of the variance function, which is proposed by the nonparametric kernel estimation. The asymptotic properties of the resulting estimators are established under some regularity conditions. Moreover, we further proposed a weighted bias-corrected profile Lagrange multiplier test statistic to check whether the linear restriction of the model is valid. Finally, some simulation studies and a real data example are conducted to assess the performance of our proposed estimators and the testing procedure in finite samples.

Robust MAVE for single-index varying-coefficient models

Zhao Yang,Yue Lili,Li Gaorong 한국통계학회 2022 Journal of the Korean Statistical Society Vol.51 No.4

In this paper, a robust, efficient and easily implemented estimation procedure for single-index varying-coefficient models is proposed by combining minimum average variance estimation (MAVE) with exponential squared loss. The merit of the proposed method is robust against outliers or heavy-tailed error distributions while asymptotically efficient as the original MAVE under the normal error case. A practical minorization–maximization algorithm is proposed for implementation. Under some regularity conditions, asymptotic distributions of the resulting estimators are derived. Simulation studies and a real data example are conducted to examine the finite sample performance of the proposed method. Both theoretical and empirical findings confirm that our proposed method works very well.

Correction: Robust MAVE for single‑index varying‑coefficient models

Zhao Yang,Yue Lili,Li Gaorong 한국통계학회 2022 Journal of the Korean Statistical Society Vol.51 No.4

In this paper, a robust, efficient and easily implemented estimation procedure for single-index varying-coefficient models is proposed by combining minimum average variance estimation (MAVE) with exponential squared loss. The merit of the proposed method is robust against outliers or heavy-tailed error distributions while asymptotically efficient as the original MAVE under the normal error case. A practical minorization–maximization algorithm is proposed for implementation. Under some regularity conditions, asymptotic distributions of the resulting estimators are derived. Simulation studies and a real data example are conducted to examine the finite sample performance of the proposed method. Both theoretical and empirical findings confirm that our proposed method works very well.

Revisiting feature selection for linear models with FDR and power guarantees

Yuan Panxu,Feng Sanying,Li Gaorong 한국통계학회 2022 Journal of the Korean Statistical Society Vol.51 No.4

The problem of feature selection for linear models is re-examined by using the fixed-X knockoff procedure and incorporating the selection probability as variable importance scores. Unlike previous work that predominantly focused on false discovery rate (FDR) control, this paper aims to establish theoretical power guarantees for the fixed-X knockoff procedure in linear models. An intersection selection procedure is proposed to make better use of sample data for estimating the selection probability, which in practice results in increasing the selection power. In addition, a two-stage procedure by using the data splitting technique is further developed to explore related theoretical results under high dimensionality. The performance of the proposal over its main competitors is demonstrated through comprehensive simulation studies and real data analysis.

Controlled synthesis and characterization of BiVO4 dendrites via a hydrothermal method

Yonggang Wang,Linlin Yang,Xiaofeng Wang,Yujiang Wang,GAORONG HAN 한양대학교 세라믹연구소 2016 Journal of Ceramic Processing Research Vol.17 No.7

BiVO4 dendrites have been controllably synthesized by a hydrothermal method without any surfactants or templates. Thestructure and morphology of the obtained BiVO4 dendrites were characterized by X-ray powder diffraction, scanning electronmicroscope, transmission electron microscopy, and high-resolution transmission electron microscopy. The effect of the pHvalues, precursors, solution concentrations, reaction temperature, and reaction time on the morphology and structure of theBiVO4 dendrites was systematically studied for the first time. It is found that the morphologies of the obtained BiVO4crystallites can vary from cubic-like to dendritic shape. The BiVO4 dendrites can be successfully fabricated by thehydrothermal method at 150 oC and pH 7 when Bi2(CO3)3 and NH4VO3 were used as precursors The resultant dendriticstructure has four trunks which have ordered branches on the opposite sides of the trunks. A rational mechanism for theoriental growth of the BiVO4 dendrites is discussed. The preparation of BiVO4 dendrites with well-dened shapes may open newopportunities for wide applications of future nanodevices.

EFFECTS OF PREPARATION PARAMETERS ON ONE-DIMENSIONAL CdS NANOSTRUCTURES BY DIPHENYLTHIOCARBAZONE-ASSISTED SOLVOTHERMAL METHOD

JING ZHOU,GAOLING ZHAO,BIN SONG,WEIXIA DONG,JINJIAN YANG,GAORONG HAN 성균관대학교(자연과학캠퍼스) 성균나노과학기술원 2012 NANO Vol.7 No.2

One-dimensional (1D) single crystalline CdS nanostructures have been successfully synthesized via a diphenylthiocarbazone-assisted solvothermal route. The results revealed that the microstructure and optical absorption properties of 1D CdS nanostructures were temperature and time dependent owing to thermodynamically and kinetically controlled growth. Single crystalline CdS nanowires with a diameter of 80 nm and length of 20 μm were synthesized at 180°C for 96 h. At moderate temperature (180°C), the morphology of the products transformed from irregular short rods to uniform long rods as the reaction time prolonging in a kinetically controlled growth regime. Only nanorods were obtained at a temperature as low as 120°C even extending reaction time to 260 h due to thermodynamic limited growth. At high temperature (250°C in this system), the products were nanowires with larger diameter but lower aspect ratio since the growth rates on both lateral and axial directions were accelerated. Moreover, the optical absorption spectra revealed that the CdS nanowires showed a blue shift compared with bulk CdS. Two optical absorption peaks appeared due to the nanometer effect in the radial and lengthwise directions of CdS nanowires. The growth mechanism of 1D CdS nanostructures was discussed.

Controlled synthesis of SrCO3 dendrites by a simple hydrothermal method

Yonggang Wang,Linlin Yang,Xiaofeng Wang,Yujiang Wang,GAORONG HAN 한양대학교 세라믹연구소 2015 Journal of Ceramic Processing Research Vol.16 No.2

In this paper, we report the controlled synthesis of SrCO3 crystals by a simple hydrothermal method without any surfactants and templates. The as-prepared samples were characterized by X-ray powder diffraction (XRD) and transmission electron microscopy (TEM). The effect of pH value, temperature, and reaction time on the formation of SrCO3 crystals was investigated. This novel route is proved to be simple and environment-friendly, which can be extended to the shape-controlled synthesis of other metal oxide nanocrystals.