http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Model averaging procedure for varying-coefficient partially linear models with missing responses
Jie Zeng,Weihu Cheng,Guozhi Hu,Yaohua Rong 한국통계학회 2018 Journal of the Korean Statistical Society Vol.47 No.3
This paper is concerned with model averaging procedure for varying-coefficient partially linear models with missing responses. The profile least-squares estimation process and inverse probability weighted method are employed to estimate regression coefficients of the partially restricted models, in which the propensity score is estimated by the covariate balancing propensity score method. The estimators of the linear parameters are shown to be asymptotically normal. Then we develop the focused information criterion, formulate the frequentist model averaging estimators and construct the corresponding confidence intervals. Some simulation studies are conducted to examine the finite sample performance of the proposed methods. We find that the covariate balancing propensity score improves the performance of the inverse probability weighted estimator. We also demonstrate the superiority of the proposed model averaging estimators over those of existing strategies in terms of mean squared error and coverage probability. Finally, our approach is further applied to a real data example.
Estimation in partially linear time-varying coefficients panel data models with fixed effects
Jing Zhao,Sanying Feng,Weihu Cheng 한국통계학회 2017 Journal of the Korean Statistical Society Vol.46 No.2
A partially time-varying coefficient time series panel data model with fixed effects is considered to characterize the nonlinearity and trending phenomenon in panel data model. To estimate the linear regression coefficient and the time-varying coefficient function, two methods are applied with the help of profile least squares. The first one is taking cross-sectional average to eliminate the fixed effects. The second one is local linear dummy variable approach. In each method we derive consistent estimates for both the parametric component and non-parametric trend function. The asymptotic distributions of the estimates are established when T and N tend to infinity simultaneously, where N is the cross section size, T is the time series length. The asymptotic results reveal that the parametric component (non-parametric coefficient function) estimate based on crosssectional average has a rate of convergence T)^−1/2 ((Th)^−1/2) that is slower than that based on local linear dummy variable approach, which is (NT))^−1/2((NTh)^−1/2), where h is the bandwidth. Furthermore, block bootstrap method is used to construct confidence interval for parametric and nonparametric components, respectively. At last, some simulation studies are conducted to examine the finite sample performance for the proposed methods and a real data example is analyzed.