Journal of Inequalities and Applications

Table 1 Biases of α and β 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	\(\boldsymbol{\hat{\alpha}_{1}}\)	\(\boldsymbol{\hat{\alpha}_{2}}\)	\(\boldsymbol{\hat{\alpha}_{3}}\)	\(\boldsymbol{\hat{\beta}}\)	\(\boldsymbol{\breve{\alpha}_{1}}\)	\(\boldsymbol{\breve{\alpha}_{2}}\)	\(\boldsymbol{\breve{\alpha}_{3}}\)	\(\boldsymbol{\breve{\beta}}\)
0.30	50	0.0035	0.0041	−0.0057	0.0308	−0.0026	0.0037	−0.0031	0.0172
	100	0.0032	−0.0018	−0.0024	0.0230	0.0016	−0.0009	−0.0010	0.0092
	150	0.0036	−0.0012	−0.0026	0.0234	0.0016	−0.0005	−0.0011	0.0093
0.20	50	0.0033	0.0023	−0.0047	0.0206	−0.0012	0.0009	−0.0008	0.0147
	100	0.0030	−0.0018	−0.0023	0.0162	0.0012	−0.0007	−0.0005	0.0089
	150	0.0031	−0.0011	−0.0024	0.0167	0.0013	−0.0003	−0.0006	0.0091
0.10	50	0.0021	0.0009	−0.0029	0.0132	−0.0005	0.0005	0.0004	0.0101
	100	0.0022	−0.0009	−0.0013	0.0097	0.0003	−0.0002	−0.0004	0.0063
	150	0.0022	−0.0009	−0.0015	0.0096	0.0004	−0.0002	−0.0005	0.0056

Back to article page