Table 9 Result on Flash loan attack
From: Blockchain-oriented approach for detecting cyber-attack transactions
Category | Algorithm | Train | Test | ||||
|---|---|---|---|---|---|---|---|
Precision | Recall | F1 score | Precision | Recall | F1 score | ||
SML | XGB | 0.934 ± 0.007 | 0.803 ± 0.086 | 0.861 ± 0.053 | 0.677 ± 0.010 | 0.557 ± 0.079 | 0.608 ± 0.044 |
RF | 0.991 ± 0.009 | 0.991 ± 0.009 | 0.991 ± 0.000 | 0.767 ± 0.045 | 0.578 ± 0.013 | 0.658 ± 0.008 | |
LGBM | 0.981 ± 0.000 | 0.944 ± 0.018 | 0.962 ± 0.010 | 0.719 ± 0.031 | 0.512 ± 0.034 | 0.598 ± 0.034 | |
UML | CBLOF | 0.825 ± 0.009 | 0.909 ± 0.000 | 0.865 ± 0.005 | 0.809 ± 0.017 | 0.974. ± 0.027 | 0.894 ± 0.021 |
HBOS | 0.893 ± 0.012 | 0.852 ± 0.011 | 0.872 ± 0.016 | 0.881 ± 0.024 | 0.928 ± 0.019 | 0.904 ± 0.022 | |
KNN | 0.857 ± 0.009 | 0.886 ± 0.000 | 0.872 ± 0.019 | 0.838 ± 0.019 | 0.947 ± 0.000 | 0.889 ± 0.011 | |
Avg KNN | 0.919 ± 0.032 | 0.886 ± 0.000 | 0.902 ± 0.005 | 0.900 ± 0.005 | 0.909 ± 0.000 | 0.904 ± 0.009 | |
LOF | 0.785 ± 0.008 | 0.500 ± 0.023 | 0.611 ± 0.019 | 0.833 ± 0.009 | 0.789 ± 0.053 | 0.810 ± 0.032 | |
OCSVM | 0.842 ± 0.009 | 0.909 ± 0.000 | 0.874 ± 0.005 | 0.845 ± 0.019 | 0.928 ± 0.019 | 0.884 ± 0.019 | |
FB | 0.628 ± 0.006 | 0.614 ± 0.023 | 0.620 ± 0.009 | 0.612 ± 0.053 | 0.789 ± 0.053 | 0.689 ± 0.022 | |
DeepSVDD | 0.860 ± 0.047 | 0.886 ± 0.000 | 0.872 ± 0.024 | 0.839 ± 0.108 | 0.974 ± 0.026 | 0.896 ± 0.051 | |
VAE | 0.833 ± 0.000 | 0.898 ± 0.012 | 0.864 ± 0.005 | 0.864 ± 0.000 | 0.947 ± 0.000 | 0.903 ± 0.000 | |
IF | 0.870 ± 0.019 | 0.909 ± 0.000 | 0.889 ± 0.010 | 0.884 ± 0.068 | 0.909 ± 0.000 | 0.895 ± 0.035 | |
EIF | 0.835 ± 0.035 | 0.909 ± 0.000 | 0.870 ± 0.019 | 0.827 ± 0.043 | 0.909 ± 0.000 | 0.865 ± 0.023 | |
WEIF | 0.880 ± 0.055 | 0.909 ± 0.000 | 0.894 ± 0.029 | 0.888 ± 0.063 | 0.928 ± 0.019 | 0.906 ± 0.045 | |