Efficient Deep Learning Models for DGA Domain Detection

Table 5

Overall classification performances of deep learning models.

Run	CNN			LSTM_Attention			BiLSTM_Attention			Ensemble
	P	R	F1	P	R	F1	P	R	F1	P	R	F1
1	0.9327	0.9383	0.9303	0.9598	0.9615	0.9591	0.9620	0.9636	0.9619	0.9670	0.9681	0.9664
2	0.9429	0.9487	0.9423	0.9607	0.9618	0.9604	0.9624	0.9635	0.9618	0.9673	0.9682	0.9664
3	0.9416	0.9475	0.9410	0.9598	0.9607	0.9590	0.9626	0.9634	0.9619	0.9677	0.9683	0.9668
4	0.9373	0.9461	0.9393	0.9605	0.9620	0.9597	0.9620	0.9636	0.9616	0.9677	0.9684	0.9666
5	0.9400	0.9462	0.9392	0.9606	0.9622	0.9604	0.9622	0.9630	0.9619	0.9680	0.9679	0.9670
Min	0.9327	0.9383	0.9303	0.9598	0.9607	0.9590	0.9620	0.9630	0.9616	0.9670	0.9679	0.9664
Max	0.9429	0.9487	0.9423	0.9607	0.9622	0.9604	0.9626	0.9636	0.9619	0.9680	0.9684	0.9670
Avg	0.9389	0.9454	0.9384	0.9603	0.9616	0.9597	0.9622	0.9634	0.9618	0.9676	0.9682	0.9666

P : Precision, R : Recall, and F1 : F1-score.