Skip to main content

Advertisement

Table 5 Absolute deviations between overall accuracies reported for un-permuted and permuted cases, by each predictor variable for the final logistic model (LR-Reduced)

From: Statistical prediction of immunity to placental malaria based on multi-assay antibody data for malarial antigens

Predictor Overall accuracy Specificity Sensitivity AUC
Age 0.001 (18) 0.001 (18) 0.002 (15) <0.001 (19)
Gravidity 0.002 (17) 0.001 (19) 0.002 (14) 0.003 (13)
FV2 FCR3 0.007 (08) 0.011 (06) 0.005 (11) 0.006 (09)
DBL2 FCR3 0.003 (13) 0.008 (10) 0.001 (18) 0.009 (06)
DBL3 FCR3 0.002 (15) 0.007 (11) 0.001 (16) 0.002 (14)
DBL4 FCR3 0.007 (07) 0.007 (12) 0.007 (08) 0.001 (17)
DBL6 FCR3 0.012 (05) 0.009 (09) 0.013 (05) 0.009 (07)
ID1-ID2a FCR3 0.005 (09) 0.011 (05) 0.004 (12) 0.001 (16)
AMA-1 0.036 (01) 0.028 (02) 0.039 (01) 0.032 (01)
CSP 0.003 (14) 0.010 (08) 0.001 (17) 0.003 (12)
EBA-175 0.017 (02) 0.018 (03) 0.017 (02) 0.019 (03)
LSA1 0.005 (11) 0.002 (17) 0.006 (10) 0.001 (18)
MSP1 0.005 (12) 0.003 (15) 0.007 (07) 0.002 (15)
MSP2 0.017 (03) 0.029 (01) 0.013 (04) 0.021 (02)
MSP3 <0.001 (19) 0.004 (13) <0.001 (19) 0.004 (11)
MSP11 0.010 (06) 0.013 (04) 0.009 (06) 0.011 (05)
Pf41 0.014 (04) 0.010 (07) 0.016 (03) 0.011 (04)
Pf70 0.002 (16) 0.004 (14) 0.002 (13) 0.004 (10)
RESA 0.005 (10) 0.003 (16) 0.006 (09) 0.007(08)
  1. Values inside brackets provide ranks of each variable. Rank one is given to the variable with the highest absolute deviation