. 2016 Feb 12;6:21689. doi: 10.1038/srep21689

Table 3. Simulation results for estimating β ₁ = β ₂ = β ₃ = 1 in the presence of missing data based on 200 monte carlo data sets, where n = 100 and ρ = 0.9.


		Bias	SE	SD	MSE	CR	Bias	SE	SD	MSE	CR	Bias	SE	SD	MSE	CR
q = 4	GS	0.029	0.515	0.476	0.226	0.960	0.003	0.512	0.521	0.270	0.945	0.008	0.517	0.489	0.238	0.970
	CC	−0.219	0.583	0.609	0.417	0.905	−0.245	0.585	0.560	0.371	0.925	−0.242	0.582	0.582	0.395	0.945
	MI-true	−0.012	0.711	0.699	0.487	0.935	−0.027	0.710	0.715	0.509	0.915	−0.007	0.701	0.685	0.467	0.940
	p = 200
	MICE-RF	−0.483	0.721	0.358	0.361	0.980	−0.277	0.704	0.376	0.218	1.000	−0.372	0.722	0.380	0.283	0.975
	KNN-V	−1.034	0.747	0.584	1.408	0.765	−0.113	0.741	0.596	0.366	0.990	−0.676	0.758	0.630	0.852	0.895
	KNN-S	−0.219	0.583	0.609	0.417	0.905	−0.245	0.585	0.560	0.371	0.925	−0.242	0.582	0.582	0.395	0.945
	MICE-DURR(Lasso)	−0.857	0.547	0.435	0.922	0.670	−0.549	0.532	0.417	0.475	0.905	−0.643	0.545	0.441	0.606	0.835
	MICE-DURR(EN)	−0.862	0.548	0.443	0.939	0.670	−0.534	0.537	0.420	0.461	0.880	−0.649	0.550	0.438	0.612	0.815
	MICE-DURR(Alasso)	−0.841	0.571	0.409	0.874	0.745	−0.525	0.564	0.377	0.417	0.935	−0.677	0.576	0.396	0.615	0.855
	MICE-IURR(Lasso)	−0.049	0.676	0.701	0.491	0.915	−0.047	0.671	0.694	0.481	0.920	−0.027	0.669	0.702	0.492	0.920
	MICE-IURR(EN)	−0.028	0.671	0.716	0.511	0.915	−0.059	0.676	0.707	0.501	0.910	−0.052	0.677	0.706	0.498	0.935
	MICE-IURR(Alasso)	−0.005	0.684	0.710	0.502	0.920	−0.024	0.685	0.727	0.526	0.915	−0.031	0.679	0.710	0.503	0.935
	p = 1000
	MICE-DURR(Lasso)	−0.787	0.556	0.399	0.778	0.740	−0.494	0.538	0.405	0.408	0.920	−0.688	0.559	0.375	0.613	0.845
	MICE-DURR(EN)	−0.793	0.554	0.414	0.799	0.715	−0.490	0.534	0.410	0.407	0.900	−0.681	0.562	0.373	0.603	0.840
	MICE-DURR(Alasso)	−0.801	0.572	0.363	0.772	0.775	−0.496	0.557	0.369	0.382	0.930	−0.705	0.577	0.339	0.611	0.880
	MICE-IURR(Lasso)	−0.014	0.682	0.744	0.550	0.930	−0.058	0.672	0.763	0.582	0.915	−0.144	0.680	0.709	0.521	0.920
	MICE-IURR(EN)	0.009	0.679	0.729	0.529	0.935	−0.062	0.667	0.744	0.555	0.915	−0.155	0.672	0.702	0.515	0.930
	MICE-IURR(Alasso)	−0.029	0.729	0.739	0.545	0.935	−0.024	0.716	0.818	0.666	0.895	−0.130	0.715	0.760	0.592	0.910
q = 20	GS	0.018	0.514	0.478	0.228	0.945	−0.010	0.512	0.519	0.268	0.945	−0.003	0.515	0.488	0.237	0.975
	CC	−0.260	0.570	0.562	0.382	0.930	−0.295	0.566	0.533	0.370	0.930	−0.304	0.568	0.537	0.379	0.925
	MI-true	−0.056	0.681	0.561	0.316	0.975	−0.075	0.671	0.590	0.352	0.980	−0.009	0.667	0.566	0.319	0.960
	p = 200
	MICE-RF	−0.334	0.747	0.361	0.241	0.995	−0.153	0.707	0.362	0.154	1.000	−0.269	0.756	0.398	0.230	1.000
	KNN-V	−0.798	0.749	0.555	0.943	0.895	−0.108	0.739	0.555	0.318	1.000	−0.506	0.757	0.569	0.578	0.960
	KNN-S	−0.260	0.570	0.562	0.382	0.930	−0.295	0.566	0.533	0.370	0.930	−0.304	0.568	0.537	0.379	0.925
	MICE-DURR(Lasso)	−0.690	0.576	0.412	0.645	0.880	−0.392	0.557	0.404	0.316	0.945	−0.575	0.574	0.406	0.495	0.930
	MICE-DURR(EN)	−0.679	0.579	0.420	0.636	0.880	−0.379	0.556	0.404	0.306	0.955	−0.591	0.572	0.414	0.520	0.920
	MICE-DURR(Alasso)	−0.668	0.616	0.382	0.592	0.915	−0.393	0.593	0.391	0.306	0.970	−0.611	0.611	0.381	0.519	0.945
	MICE-IURR(Lasso)	−0.051	0.665	0.646	0.418	0.935	−0.031	0.666	0.678	0.458	0.910	0.007	0.660	0.658	0.431	0.950
	MICE-IURR(EN)	−0.031	0.665	0.645	0.415	0.935	−0.057	0.662	0.660	0.436	0.930	0.006	0.659	0.655	0.427	0.945
	MICE-IURR(Alasso)	−0.035	0.668	0.691	0.476	0.915	−0.063	0.667	0.660	0.437	0.935	−0.003	0.667	0.684	0.466	0.935
	p = 1000
	MICE-DURR(Lasso)	−0.738	0.582	0.417	0.717	0.850	−0.408	0.569	0.417	0.340	0.955	−0.531	0.586	0.444	0.478	0.905
	MICE-DURR(EN)	−0.729	0.587	0.416	0.704	0.860	−0.416	0.569	0.421	0.350	0.945	−0.526	0.587	0.443	0.472	0.925
	MICE-DURR(Alasso)	−0.723	0.614	0.384	0.670	0.920	−0.433	0.600	0.374	0.327	0.950	−0.540	0.612	0.401	0.451	0.945
	MICE-IURR(Lasso)	0.055	0.687	0.725	0.526	0.905	−0.057	0.676	0.770	0.593	0.890	−0.124	0.676	0.735	0.553	0.895
	MICE-IURR(EN)	0.052	0.675	0.746	0.556	0.900	−0.053	0.673	0.782	0.611	0.910	−0.113	0.669	0.751	0.573	0.915
	MICE-IURR(Alasso)	−0.003	0.742	0.769	0.589	0.930	−0.105	0.728	0.803	0.653	0.920	−0.082	0.732	0.792	0.631	0.915

Bias, mean bias; SE, mean standard error; SD, Monte Carlo standard deviation; MSE, mean square error; CR, coverage rate of 95% confidence interval; GS, gold standard; CC, complete-case; KNN-V, KNN by nearest variables; KNN-S, KNN by nearest subjects; MICE-DURR, MICE through direct use of regularized regressions; MICE-IURR, MICE through indirect use of regularized regressions; EN, elastic net; Alasso, adaptive lasso.

Table 3. Simulation results for estimating β 1 = β 2 = β 3 = 1 in the presence of missing data based on 200 monte carlo data sets, where n = 100 and ρ = 0.9.

Table 3. Simulation results for estimating β ₁ = β ₂ = β ₃ = 1 in the presence of missing data based on 200 monte carlo data sets, where n = 100 and ρ = 0.9.