Skip to main content
. Author manuscript; available in PMC: 2024 Jan 30.
Published in final edited form as: Pac Symp Biocomput. 2024;29:96–107.

Table 2.

Comparison of study design performance as measured by AUROC for each distance measure. Bolded metric values are significantly better than the others.

Cosine
Euclidean
Gower
Manhattan
Design* mean 95% CI mean 95% CI mean 95% CI mean 95% CI
A 0.800 0.792, 0.808 0.807 0.800, 0.814 0.791 0.781, 0.800 0.800 0.792, 0.807
B 0.793 0.785, 0.801 0.799 0.792, 0.806 0.784 0.774, 0.794 0.792 0.784, 0.800
C 0.793 0.785, 0.801 0.798 0.791, 0.805 0.783 0.773, 0.793 0.791 0.783, 0.798
D 0.840 0.833, 0.847 0.848 0.842, 0.854 0.864 0.857, 0.872 0.852 0.845, 0.858
E 0.840 0.833, 0.847 0.845 0.839, 0.852 0.852 0.844, 0.860 0.846 0.839, 0.852
F 0.510 0.500, 0.521 0.512 0.503, 0.522 0.494 0.482, 0.507 0.485 0.475, 0.495
*

Real patients (A), digital twins (B), real patients and digital twins (C), closest digital twins (D), real patients and closest digital twins (E), and real patients outside the community (F).