Research Article
Network Intrusion Detection Based on Sparse Autoencoder and IGA-BP Network
Table 8
The accuracy result of various intrusion types (%).
| Methods | Normal | DoS | Probe | R2L | U2R |
| BPNN [31] | 96.24 | 97.51 | 91.18 | 23.22 | 64.28 | PCA-PSO-BP [8] | 97.21 | 98.52 | 92.14 | 26.52 | 65.44 | PCA-ANN [6] | 96.52 | 97.07 | 91.13 | 27.63 | 68.63 | KPCA-SVM [9] | 98.56 | 98.51 | 93.11 | 26.72 | 71.28 | RNN [32] | 98.73 | 98.69 | 93.26 | 28.71 | 73.51 | DBN [33] | 99.03 | 98.71 | 94.22 | 32.56 | 75.97 | CNN-WDLSTM [34] | 99.12 | 98.97 | 99.10 | 54.25 | 78.64 | Proposed | 99.51 | 99.27 | 99.38 | 61.21 | 81.94 |
|
|