Table 5.

Evaluation indexes of different models.

Networks		GunPoint	Lightning7	Trace	OSULeaf	SyntheticControl	CBF	UMD
LSTM	Precision	0.9744	0.5908	0.2621	0.5629	0.9807	0.9879	0.8197
	Recall	0.9737	0.5486	0.5	0.5783	0.98	0.9878	0.7986
	F1-score	0.9704	0.5689	0.3811	0.5705	0.9803	0.9879	0.809
GRU	Precision	0.9602	0.746	1	0.6073	0.9901	0.9913	0.9867
	Recall	0.9602	0.7096	1	0.6094	0.99	0.9911	0.9861
	F1-score	0.9602	0.7274	1	0.6083	0.9901	0.9912	0.9864
BiLSTM	Precision	0.9744	0.6239	0.7545	0.5882	0.9803	0.9802	1
	Recall	0.9737	0.4313	0.7267	0.5781	0.98	0.9801	1
	F1-score	0.974	0.51	0.7403	0.5831	0.9802	0.9802	1
BLS + GADF	Precision	0.9735	0.6294	1	0.6018	0.9747	0.9541	0.8167
	Recall	0.9735	0.6048	1	0.5897	0.9733	0.9514	0.8194
	F1-score	0.9735	0.6168	1	0.5957	0.9740	0.9527	0.8178
BLS + RP	Precision	0.9365	0.5456	0.9565	0.6088	0.9809	0.9727	0.9723
	Recall	0.9328	0.5662	0.9643	0.5943	0.9800	0.9724	0.9722
	F1-score	0.9346	0.5557	0.9604	0.6015	0.9804	0.9725	0.9723
Best D–S	Precision	0.9805	0.7950	1	0.6857	0.9967	0.9944	1
	Recall	0.9803	0.6709	1	0.6702	0.9967	0.9945	1
	F1-score	0.9804	0.7277	1	0.6779	0.9967	0.9945	1