A Deep Learning-Based Framework for Feature Extraction and Classification of Intrusion Detection in Networks

Table 6

Comparison of the proposed work with different models.

Model	Dataset	Classifier	AC (%)	Precision	Recall	F1 score
Tang et al. [46]	NSL-KDD	SAAE-DNN	87.74	86.47	84.12	85
Wang et al. [5]	NSL-KDD	RNN	94.19	—	—	—
Al-Qatf et al. [6]	NSL-KDD	STL-IDS	84.96	96.2	76.5	85.2
Ingre et al. [7]	NSL-KDD	ANN	81.2	—	—	—
Tang et al. [8]	NSL-KDD	SDN-DNN	75.75	83	75	74
Yin et al. [9]	NSL-KDD	RNN-IDS	83.28	—	97.09	—
Li et al. [10]	NSL-KDD	GoogLeNet	81.84	81.84	100	90.01
Tama et al. [11]	NSL-KDD UNSW-NB15	TSE-IDS	85.79	88	86.80	87.4
Choudhary et al. [12]	NSL-KDD	DNN	91.7	93.6	92	92.2
Farahnakian et al. [13]	KDD99	DAE	96.53	—	—	—
YU et al. [34]	NSL-KDD UNSW-NB15	CNN	92.33	96.1	95	93
Wang et al. [48]	NSL-KDD	SDAE-ELM1	78.04	95.99	64.12	76.87
Proposed model	NSL-KDD	DNN	99.73	99.75	99.73	99.72