Table 5.

Comparison of Mean-Absolute-Error (MAE) and Rooted-Mean-Squared-Error (RMSE) for Model 1 with different link functions and the censoring distribution is mis-specificed with α = 1.5.

p	SRF	Naive.Cox	Naive.km	Lu.id	Lu.exp	Wang.id	Wang.exp
Model 1: identity link, n = 3, 000, SNR = 0.3
5	0.1363	0.1378	0.2067	0.1352	0.1357	0.1352	0.1357
	0.1701	0.1702	0.2467	0.1697	0.1702	0.1697	0.1702
10	0.1376	0.1385	0.2073	0.1358	0.1363	0.1358	0.1363
	0.1709	0.1706	0.2472	0.1699	0.1704	0.1699	0.1704
20	0.1371	0.1371	0.2062	0.1341	0.1347	0.1342	0.1347
	0.1698	0.1691	0.2464	0.1682	0.1688	0.1682	0.1688
Model 1: log-exp link, n = 3, 000, SNR = 0.3
5	0.1350	0.1366	0.2046	0.1340	0.1345	0.1340	0.1345
	0.1686	0.1687	0.2441	0.1683	0.1688	0.1683	0.1688
10	0.1363	0.1373	0.2053	0.1346	0.1352	0.1347	0.1352
	0.1695	0.1692	0.2447	0.1685	0.1690	0.1685	0.1690
20	0.1359	0.1359	0.2043	0.1330	0.1335	0.1330	0.1336
	0.1683	0.1677	0.2439	0.1669	0.1674	0.1669	0.1674
Model 1: exp link, n = 3, 000, SNR = 0.3
5	24.537	25.171	33.190	24.322	24.601	24.304	24.600
	30.701	30.750	38.999	30.549	30.735	30.532	30.715
10	24.802	25.317	33.359	24.468	24.743	24.445	24.744
	30.798	30.832	39.142	30.577	30.757	30.560	30.742
20	24.852	25.188	33.406	24.300	24.567	24.272	24.570
	30.732	30.654	39.103	30.384	30.583	30.371	30.576

The number of covariates p = 5, 10, 20, for each p, the first row is MAE, the second row is RMSE. SRF, proposed random forest-bases estimator; Naive.km, estimate based on Kaplan–Meier estimator without adjusting for the covariates; Naive.Cox, Cox regression based estimator; Lu.id, method of Tian et al. (2014) with identity link; Lu.exp, method of Tian et al. (2014) with exponential link; Wang.id, method of Wang and Schaubel (2018) with identity link; Wang:exp, method of Wang and Schaubel (2018) with exponential link.