Skip to main content

Table 1 Sensitivity for the most competitive classifiers and imputation combinations and CF

From: A comparative study: classification vs. user-based collaborative filtering for clinical prediction

Dataset (% missing) RF-MI RF-MICE LR-MI LR-MICE CF
NHANES (baseline=3%) 0.99 0.99 0.99 0.99 0.56
NHANES (10%) 0.98 0.99 0.98 0.99 0.63
SUPPORT (baseline = 16%) 0.99 0.98 0.98 0.97 0.98
Kidney (baseline =10%) 0.99 0.98 0.97 0.99 0.99
NHANES (20%) 0.98 0.98 0.99 0.99 0.49
SUPPORT (20%) 0.96 0.96 0.84 0.98 0.86
Kidney (20%) 0.98 0.99 0.98 0.98 0.98
NHANES (30%) 0.89 0.99 0.89 0.95 0.25
SUPPORT (30%) 0.86 0.99 0.81 0.93 0.82
Kidney (30%) 0.98 0.98 0.97 0.96 0.92