Table 1 Mean and standard variance of AUROC (%) values of considered algorithms

From: Realistic fault detection of li-ion battery via dynamical deep learning

Algorithm

AUROC (%)

Average Direct Cost (104 CNY)

DyAD

88.6 ± 2.9

0.085

GDN

70.3 ± 5.5

0.126

AE

72.8 ± 13.4

0.133

SVDD

51.5 ± 8.26

0.152

GP

66.6

0.162

VE

55.6

0.169

  1. Our proposed method is highlighted in bold. The average direct costs of fault and inspection are in the unit of 104 CNY/vehicle/year and calculated using the mean of the cost ranges. The VE and GP method has little internal randomness and hence no variance is reported.
Back to article page