Skip to main content

Table 3 The relative mean integrated square error of \(\pmb{g(\cdot)}\) under different missing functions and different sample sizes obtained by two different methods for the simulated data

From: Estimation in a partially linear single-index model with missing response variables and error-prone covariates

  

Complete

Imputation

Missing rate

n

\(\boldsymbol{\hat{g}_{n}(\cdot)}\)

\(\boldsymbol{\breve{g}_{n}(\cdot)}\)

0.30

50

0.2014

0.1227

 

100

0.1251

0.0921

 

150

0.1014

0.0915

0.20

50

0.1514

0.0930

 

100

0.1102

0.0915

 

150

0.1026

0.0918

0.10

50

0.1451

0.0923

 

100

0.1138

0.0910

 

150

0.0996

0.0906