Blockchain-oriented approach for detecting cyber-attack transactions

Financial Innovation

Table 8 Result on Smart contract exploit

Category	Algorithm	Train			Test
Category	Algorithm	Precision	Recall	F1 score	Precision	Recall	F1 score
SML	XGB	0.924 ± 0.039	0.572 ± 0.047	0.707 ± 0.047	0.653 ± 0.097	0.308 ± 0.045	0.419 ± 0.061
	RF	1.000 ± 0.000	0.924 ± 0.043	0.960 ± 0.023	0.633 ± 0.033	0.393 ± 0.077	0.483 ± 0.069
	LGBM	0.987 ± 0.013	0.813 ± 0.068	0.890 ± 0.035	0.619 ± 0.048	0.341 ± 0.130	0.430 ± 0.122
UML	CBLOF	0.830 ± 0.010	0.902 ± 0.011	0.865 ± 0.010	0.795 ± 0.023	0.875 ± 0.025	0.833 ± 0.024
	HBOS	0.882 ± 0.013	0.728 ± 0.011	0.798 ± 0.012	0.861 ± 0.028	0.775 ± 0.025	0.816 ± 0.026
	KNN	0.862 ± 0.011	0.880 ± 0.011	0.871 ± 0.011	0.833 ± 0.024	0.875 ± 0.025	0.854 ± 0.024
	Avg KNN	0.886 ± 0.032	0.848 ± 0.000	0.862 ± 0.015	0.871 ± 0.071	0.800 ± 0.000	0.832 ± 0.032
	LOF	0.846 ± 0.000	0.717 ± 0.000	0.776 ± 0.000	0.813 ± 0.000	0.650 ± 0.000	0.722 ± 0.000
	OCSVM	0.847 ± 0.007	0.902 ± 0.011	0.874 ± 0.001	0.834 ± 0.016	0.875 ± 0.025	0.854 ± 0.004
	FB	0.683 ± 0.017	0.750 ± 0.011	0.715 ± 0.014	0.582 ± 0.058	0.675 ± 0.125	0.624 ± 0.087
	DeepSVDD	0.866 ± 0.043	0.891 ± 0.022	0.877 ± 0.011	0.828 ± 0.120	0.875 ± 0.025	0.848 ± 0.075
	VAE	0.837 ± 0.000	0.891 ± 0.000	0.863 ± 0.000	0.850 ± 0.000	0.850 ± 0.000	0.850 ± 0.000
	IF	0.868 ± 0.021	0.859 ± 0.011	0.863 ± 0.016	0.884 ± 0.067	0.859 ± 0.011	0.869 ± 0.027
	EIF	0.837 ± 0.033	0.880 ± 0.011	0.857 ± 0.012	0.830 ± 0.037	0.880 ± 0.033	0.853 ± 0.004
	WEIF	0.875 ± 0.071	0.880 ± 0.022	0.876 ± 0.032	0.861 ± 0.089	0.900 ± 0.050	0.880 ± 0.070

Each metric is formed with its mean value and corresponding range. The corresponding range of each metric is the max difference between the mean value and its real value of the metric, hereafter