Skip to main content
. 2022 Apr 7;22(8):2837. doi: 10.3390/s22082837

Table 4.

Maintenance event detection results in Use Case 1. The best results for each metric are highlighted in bold.

Component Algorithm Sensitivity FP Rate Accuracy Distribution Threshold Mean Ratio
Comp1 MEDEP 0.975 0.948 0.051 False False
MEDEP 0.975 0.768 0.231 True False
MEDEP 0.878 0.700 0.300 False True
MEDEP 0.878 0.052 0.947 True True
LOF 0.531 0.545 0.469 - -
Comp2 MEDEP 0.943 0.936 0.063 False False
MEDEP 0.943 0.734 0.265 True False
MEDEP 0.943 0.572 0.472 False True
MEDEP 0.943 0.122 0.877 True True
LOF 0.467 0.527 0.473 - -
Comp3 MEDEP 0.900 0.963 0.037 False False
MEDEP 0.900 0.858 0.142 True False
MEDEP 0.850 0.767 0.232 False True
MEDEP 0.850 0.105 0.895 True True
LOF 0.522 0.544 0.456 - -
Comp4 MEDEP 1.000 0.908 0.092 False False
MEDEP 1.000 0.593 0.407 True False
MEDEP 0.945 0.313 0.687 False True
MEDEP 0.945 0.054 0.946 True True
LOF 0.407 0.461 0.539 - -