“…In addition, in order to demonstrate the superiority of the proposed SAVAER-DNN, the performance of SAVAER-DNN is compared with other state-of-the-art intrusion detection models reported in the intrusion detection literature, including S-NDAE (stacked nonsymmetric deep auto-encoders) [7], SCDNN (spectral clustering and deep neural network) [19], ID-CVAE (a unsupervised network intrusion detection method based on a conditional variational auto-encoder) [20], RNN-IDS (recurrent neural network) [21], ResNet50 [22], GoogLeNe [22], LSTM 4 [24], GRU 3 [24], CFBLS (BLS with cascades of mapped features) [24], SHIA (scale-hybrid-IDS-AlertNet) [25], Gaussian-Bernoulli RBM [51], Random tree+NBTree [52], TSE-IDS (Two-Stage Classifier Ensemble for IDS) [53], EM Clustering [16], DT (decision tree) [16], TSDL (two-stage deep learning model) [23], CASCADE-ANN (a multiclass cascade of artificial neural network) [54] and AODE (average one dependence estimator algorithm) [55]. To be fair, only detection models using the same test dataset are selected.…”