Estimating misclassification error: a closer look at cross-validation based methods

BMC Research Notes

Table 2 Simulation results for p = 1, Σ₁ = Σ₂ = I₍₁₎, N = 1000

Method	n	_Δ	ē ^a	ē _N	$\overset{―}{M S E}$	SD(MSE)	$\overset{―}{V A R}$	SD(VAR)	MSB	$\overset{―}{B I A S}$	SD(BIAS)
LOOCV	50	1	0.30907	0.31124	0.00446	0.00635	0	0	0.00446	0.00217	0.06677
	50	3	0.07185	0.07114	0.00130	0.00184	0	0	0.00130	−0.00071	0.03603
BCVn	50	1	0.30907	0.30791	0.00942	0.00582	0.00543	0.00204	0.00399	−0.00116	0.06317
	50	3	0.07185	0.06781	0.00256	0.00180	0.00143	0.00074	0.00113	−0.00404	0.03340
LOOCV	100	1	0.30483	0.30579	0.00191	0.00281	0	0	0.00191	0.00096	0.04366
	100	3	0.07020	0.06987	0.00075	0.00114	0	0	0.00075	−0.00033	0.02741
BCVn	100	1	0.30483	0.30469	0.00419	0.00289	0.00243	0.00074	0.00175	−0.00013	0.04188
	100	3	0.07020	0.06829	0.00139	0.00109	0.00070	0.00030	0.00069	−0.00190	0.02621
k CV10	50	1	0.30907	0.31241	0.00461	0.00625	0.00029	0.00023	0.00432	0.00334	0.06568
	50	3	0.07185	0.07123	0.00131	0.00180	0.00008	0.00008	0.00123	−0.00062	0.03511
BCV10	50	1	0.30907	0.30906	0.00915	0.00526	0.00536	0.00145	0.00379	−0.00001	0.06163
	50	3	0.07185	0.06812	0.00252	0.00173	0.00141	0.00067	0.00111	−0.00373	0.03306
k CV10	100	1	0.30483	0.30642	0.00195	0.00280	0.00009	0.00005	0.00186	0.00160	0.04307
	100	3	0.07020	0.06993	0.00075	0.00108	0.00003	0.00002	0.00072	−0.00026	0.02683
BCV10	100	1	0.30483	0.30511	0.00411	0.00254	0.00241	0.00038	0.00170	0.00028	0.04125
	100	3	0.07020	0.06848	0.00136	0.00101	0.00069	0.00025	0.00067	−0.00171	0.02577

$\overset{―}{e} a = (1 / N) \sum_{i = 1}^{N} e_{i}, {\overset{―}{e}}_{N} = (1 / N) \sum_{i = 1}^{N} {\overset{―}{e}}_{Ri}, M S B = (1 / N) \sum_{i = 1}^{N} {({\overset{―}{e}}_{Ri} - e_{i})}^{2}$

ISSN: 1756-0500