CTRF: Ethereum-Based Ponzi Contract Identification

Table 2

The experimental results on four datasets.

Metric		KNN	CNN	DT	SVM	XGBoost	CTRF
Precision	D1-code	0.518	0.486	0.500	0.586	0.918	0.953
	D2-code	0.501	0.630	0.551	0.543	0.926	0.933
	D1-codeAndTran	0.485	0.621	0.571	0.642	0.907	0.929
	D2-codeAndTran	0.478	0.705	0.611	0.545	0.918	0.928
Recall	D1-code	0.803	0.583	0.836	0.836	0.811	0.847
	D2-code	0.813	0.773	0.812	0.788	0.863	0.875
	D1-codeAndTran	0.787	0.683	0.721	0.852	0.828	0.852
	D2-codeAndTran	0.801	0.761	0.863	0.763	0.873	0.891
F-score	D1-code	0.628	0.530	0.626	0.689	0.862	0.897
	D2-code	0.619	0.694	0.657	0.643	0.893	0.903
	D1-codeAndTran	0.601	0.651	0.638	0.732	0.865	0.889
	D2-codeAndTran	0.598	0.732	0.715	0.635	0.894	0.909