Table 3 Generalization performance on T1D/T2D benchmark data

Method	ACC	NMI	ARI	CI
k-means+Cox PH	0.40 ± 0.01*(0.008)	0.06 ± 0.01*(0.008)	0.09 ± 0.01*(0.008)	0.51 ± 0.01*(0.008)
SSC	0.40 ± 0.01*(0.008)	0.06 ± 0.02*(0.008)	0.087 ± 0.009*(0.008)
SCA	0.53 ± 0.03*(0.008)	0.005 ± 0.006*(0.008)	0.02 ± 0.02*(0.008)	0.60 ± 0.02*(0.008)
DSM	0.498 ± 0.007*(0.011)	0.0003 ± 0.0005*(0.011)	0.003 ± 0.009*(0.011)	0.37 ± 0.02*(0.008)
RDSM	0.503 ± 0.003*(0.011)	0.002 ± 0.002*(0.011)	0.006 ± 0.008*(0.011)	0.63 ± 0.13*(0.008)
VaDeSC-MLP	0.71 ± 0.09*(0.016)	0.15 ± 0.09(0.222)	0.13 ± 0.14(0.548)	0.72 ± 0.07(0.690)
VaDeSC-EHR_nosurv	0.64 ± 0.06*(0.008)	0.06 ± 0.05*(0.008)	0.08 ± 0.03*(0.008)
VaDeSC-EHR_relage	0.81 ± 0.02(0.458)	0.24 ± 0.04(0.458)	0.23 ± 0.04(0.458)	0.71 ± 0.01*(0.047)
VaDeSC-EHR	0.81 ± 0.02	0.24 ± 0.04	0.23 ± 0.04	0.72 ± 0.01

Comparison between VaDeSC-EHR and the other methods used for clustering longitudinal survival data, in terms of balanced accuracy (ACC), normalized mutual information (NMI), adjusted Rand index (ARI), and concordance index (CI). Reported ± is one standard deviation, and the significance of any difference (p-value < 0.05) between VaDeSC-EHR and the other methods is indicated by an asterisk. Significance is based on a two-sided Mann–Whitney U-test. Detailed p-values are shown in the brackets.

Quick links

Search