A Novel Machine Learning-Based Analysis Model for Smart Contract Vulnerability

Table 5

The result of experiment.

Vulnerabilities	Models						Tools
Names	KNN			SGD			Oyente			SmartCheck
	Accuracy (%)	Recall (%)	Precision (%)	Accuracy (%)	Recall (%)	Precision (%)	Accuracy (%)	Recall (%)	Precision (%)	Accuracy (%)	Recall (%)	Precision (%)
Re-entrancy	95.45	95.45	95.83	95.45	68.18	70.83	70.10	72.10	72.10	75.20	60.10	63.50
Arithmetic	95.54	90.90	91.10	90.10	98.10	95.10	74.50	62.10	64.50	85.20	53.10	65.54
Access control	95.00	95.00	96.67	95.45	94.44	95.54	50.10	54.10	52.50	67.20	53.10	65.54
Bad randomness	90.10	93.10	91.10	74.10	75.10	73.10	46.10	47.20	47.30	54.10	55.10	56.10
Denial of service	90.20	90.26	90.56	64.20	65.10	62.40	47.10	48.42	43.64	61.10	59.10	60.20
Unchecked low level	90.10	90.10	90.10	81.10	81.20	80.10	65.65	63.34	54.24	59.10	50.20	50.10
Short address	91.10	91.10	92.85	78.10	77.27	82.14	49.30	48.45	47.65	67.10	61.10	61.10
Front running	95.45	95.45	96.10	89.90	86.10	86.20	54.27	56.18	59.26	55.20	56.20	54.10