Skip to main content

Table 5 Prediction performance of recalibrated lightgbm under CN220 guideline with different combinations of GV

From: Ensemble learning for the early prediction of neonatal jaundice with genetic features

Variables AUC F1-score Precision Specific GV
CRF 0.795 (0.761–830) 0.217 (0.171–0.261) 0.143 (0.113–0.171) None
CRF + GV4 0.807 (0.779–841) 0.242 (0.195–0.286) 0.165 (0.132–0.194) HMOX1 (1)
 rs2071749
UGT1A1 (3):
rs4148323, rs6717546, rs6719561
CRF + GV11 0.813 (0.781–0.847) 0.251 (0.198–0.298) 0.176 (0.141–0.207) HMOX1 (3)
 (GT)n, rs9607267, rs2071749
UGT1A1 (8)
 rs887829, (TA)n, rs4148323, rs1018124, rs6717546, rs11563250, rs6719561, rs4663972
CRF + GV36 0.830 (0.802–0.826) 0.285 (0.229–0.333) 0.217 (0.173–0.252) HMOX1 (4)
 rs2071746, (GT)n, rs9607267, rs2071749
UGT1A1 (9)
 rs4399719, rs887829, (TA)n, rs4148323, rs1018124, rs6717546, rs11563250, rs6719561, rs4663972
BLVRA (6)
 rs1181601, rs1181574, rs10486752, rs699512, rs17246016, rs589570
SLCO1B1 (17)
 rs4149013, rs10743408, rs3899743, rs981262, rs7138177, rs4149026, rs976754, rs4149034, rs12313639, rs2306283, rs4149044, rs4149056, rs4149057, rs4363657, rs4149076, rs12578392, rs4149085
  1. 95% CI is shown in parentheses