VdaBSC: A Novel Vulnerability Detection Approach for Blockchain Smart Contract by Dynamic Analysis

Table 6

Comparing the existing methods with the proposed model on 90% training data.

Models	Measures	Reentrancy	Integer overflow	Integer underflow	TOD	URV
Oyente	F1-measure	88.21	88.84	88.97	89.06	89.18
Oyente	Accuracy	98.63	99.26	99.39	99.48	99.6
Oyente	Precision	96.26	96.89	97.02	97.11	97.23
Oyente	Recall	92.77	93.4	93.53	93.62	93.74
Maian	F1-measure	90.21	90.84	90.97	91.06	91.18
Maian	Accuracy	98.62	99.25	99.38	99.47	99.59
Maian	Precision	96.53	97.16	97.29	97.38	97.5
Maian	Recall	93.53	94.16	94.29	94.38	94.5
SmartCheck	F1-measure	90.16	90.79	90.92	91.01	91.13
SmartCheck	Accuracy	98.68	99.31	99.44	99.53	99.65
SmartCheck	Precision	96.64	97.27	97.4	97.49	97.61
SmartCheck	Recall	93.69	94.32	94.45	94.54	94.66
ContractGuard	F1-measure	91.45	92.08	92.21	92.3	92.42
ContractGuard	Accuracy	98.56	99.19	99.32	99.41	99.53
ContractGuard	Precision	97.03	97.66	97.79	97.88	98
ContractGuard	Recall	94.38	95.01	95.14	95.23	95.35
ContractFuzzer	F1-measure	91.45	92.08	92.21	92.3	92.42
ContractFuzzer	Accuracy	98.7	99.33	99.46	99.55	99.67
ContractFuzzer	Precision	97.01	97.64	97.77	97.86	97.98
ContractFuzzer	Recall	94.41	95.04	95.17	95.26	95.38
Proposed	F1-measure	94.37	95	95.13	95.22	95.34
Proposed	Accuracy	98.98	99.61	99.74	99.83	99.95
Proposed	Precision	98.23	98.86	98.99	99.08	99.2
Proposed	Recall	96.18	96.81	96.94	97.03	97.15