I²DS: Interpretable Intrusion Detection System Using Autoencoder and Additive Tree

Table 7

Comparison of the proposed model and other classifiers with UNSW-NB15.

Classifier	Accuracy	Precision	Recall	F1-score
IGAN-IDS [9]	0.8253	0.8486	0.8445	0.8286
NADS-RA [10]	0.9490	0.9820	0.9250	0.9530
CWGAN-CSSAE [11]	0.9327	0.9259	0.9543	0.9399
HNGFA [21]	0.9024	0.9277	0.8248	0.8536
RHF-ANN [22]	0.9760	0.9550	0.9990	0.9770
GA + SVM [23]	0.9610	0.9830	0.9820	0.9820
SVM + EML + K-means [24]	0.9450	0.9480	0.9970	0.9720
Wrapper + neurotree [25]	0.9710	0.9500	0.9830	0.9660
HC-IBGSA + SVM [26]	0.9847	—	—	—
MINDFUL [27]	0.9340	—	—	0.9529
I^2DS (our proposed model)	0.9995	0.9994	0.9999	0.9996