TY - JOUR
T1 - Transformation-based estimation
AU - Feng, Zhenghui
AU - Wang, Tao
AU - ZHU, Lixing
N1 - Funding Information:
Lixing Zhu was supported by a grant from the Research Grants Council of Hong Kong and a Faculty Research Grant (FRG) from Hong Kong Baptist University . Feng’s research was supported by a National Natural Science Foundation of China grant ( # 11301434 ). The authors also thank the editor and two referees for their constructive suggestions and comments which led to a great improvement of an early manuscript.
PY - 2014/10
Y1 - 2014/10
N2 - To alleviate the computational burden of making the relevant estimation algorithms stable for nonlinear and semiparametric regression models with, particularly, high-dimensional data, a transformation-based method combining sufficient dimension reduction approach is proposed. To this end, model-independent transformations are introduced to models under study. This generic methodology can be applied to transformation models; generalized linear models; and their corresponding quantile regression variants. The constructed estimates almost have closed forms in certain sense such that the above goals can be achieved. Simulation results show that, in finite sample cases with high-dimensional predictors and long-tailed distributions of error, the new estimates often exhibit a smaller degree of variance, and have much less computational burden than the classical methods such as the classical least squares and quantile regression estimation.
AB - To alleviate the computational burden of making the relevant estimation algorithms stable for nonlinear and semiparametric regression models with, particularly, high-dimensional data, a transformation-based method combining sufficient dimension reduction approach is proposed. To this end, model-independent transformations are introduced to models under study. This generic methodology can be applied to transformation models; generalized linear models; and their corresponding quantile regression variants. The constructed estimates almost have closed forms in certain sense such that the above goals can be achieved. Simulation results show that, in finite sample cases with high-dimensional predictors and long-tailed distributions of error, the new estimates often exhibit a smaller degree of variance, and have much less computational burden than the classical methods such as the classical least squares and quantile regression estimation.
KW - Generalized linear model
KW - Model-independent transformation
KW - Quantile regression
KW - Transformation model
UR - http://www.scopus.com/inward/record.url?scp=84901035204&partnerID=8YFLogxK
U2 - 10.1016/j.csda.2014.05.001
DO - 10.1016/j.csda.2014.05.001
M3 - Journal article
AN - SCOPUS:84901035204
SN - 0167-9473
VL - 78
SP - 186
EP - 205
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
ER -