A mixed model approach to measurement error in semiparametric regression

An essential assumption in traditional regression techniques is that predictors are measured without errors. Failing to take into account measurement error in predictors may result in severely biased inferences. Correcting measurement-error bias is an extremely difficult problem when estimating a regression function nonparametrically. We propose an approach to deal with measurement errors in predictors when modelling flexible regression functions. This approach depends on directly modelling the mean and the variance of the response variable after integrating out the true unobserved predictors in a penalized splines model. We demonstrate through simulation studies that our approach provides satisfactory prediction accuracy largely outperforming previously suggested local polynomial estimators even when the model is incorrectly specified and is competitive with the Bayesian estimator.

[1]  David Ruppert,et al.  Semiparametric Regression with R , 2018 .

[2]  Jianqing Fan,et al.  Nonparametric regression with errors in variables , 1993 .

[3]  Paul H. C. Eilers,et al.  Flexible smoothing with B-splines and penalties , 1996 .

[4]  David Ruppert,et al.  Local polynomial regression and simulation–extrapolation , 2004 .

[5]  Adelchi Azzalini,et al.  The Skew-Normal and Related Families , 2018 .

[6]  Haiming Zhou,et al.  An alternative local polynomial estimator for the error-in-variables problem , 2017, 1701.06105.

[7]  Roger Logan,et al.  Estimation and Inference for Logistic Regression with Covariate Misclassification and Measurement Error in Main Study/Validation Study Designs , 2000 .

[8]  B. Ripley,et al.  Semiparametric Regression: Preface , 2003 .

[9]  Andrew W. Roddam,et al.  Measurement Error in Nonlinear Models: a Modern Perspective , 2008 .

[10]  Aurore Delaigle,et al.  Using SIMEX for Smoothing-Parameter Choice in Errors-in-Variables Problems , 2008 .

[11]  Stephen Senn,et al.  Covariance analysis in generalized linear measurement error models. , 1989, Statistics in medicine.

[12]  David Ruppert,et al.  On the asymptotics of penalized spline smoothing , 2011 .

[13]  Alexander Meister,et al.  Nonparametric Regression with Errors-in-Variables , 2009 .

[14]  R. Carroll,et al.  Bayesian semiparametric regression in the presence of conditionally heteroscedastic measurement and regression errors , 2014, Biometrics.

[15]  S Senn.  Covariance analysis in generalized linear measurement error models. , 1990, Statistics in medicine.

[16]  Irène Gijbels,et al.  Practical bandwidth selection in deconvolution kernel density estimation , 2004, Comput. Stat. Data Anal..

[17]  L. Skovgaard.  NONLINEAR MODELS FOR REPEATED MEASUREMENT DATA. , 1996 .

[18]  M. Wand,et al.  ADDITIVE MODELS WITH PREDICTORS SUBJECT TO MEASUREMENT ERROR , 2005 .

[19]  J. R. Cook,et al.  Simulation-Extrapolation Estimation in Parametric Measurement Error Models , 1994 .

[20]  A. Azzalini.  A class of distributions which includes the normal ones , 1985 .

[21]  ohn,et al.  Additive models with predictors subject to measurement error , 2004 .

[22]  Estimation of integrated squared density derivatives from a contaminated sample , 2002 .

[23]  David Ruppert,et al.  Theory & Methods: Spatially‐adaptive Penalties for Spline Fitting , 2000 .

[24]  Alexander Kukush,et al.  Measurement Error Models , 2011, International Encyclopedia of Statistical Science.

[25]  Raymond J. Carroll,et al.  Measurement error in nonlinear models: a modern perspective , 2006 .

[26]  Scott M. Berry,et al.  Bayesian Smoothing and Regression Splines for Measurement Error Problems , 2002 .

[27]  Raymond J Carroll,et al.  A Design-Adaptive Local Polynomial Estimator for the Errors-in-Variables Problem , 2009, Journal of the American Statistical Association.

[28]  D. Ruppert,et al.  Nonparametric regression in the presence of measurement error , 1999 .