Table 8.
Performance comparison of binary classification. The best results for the models are extracted from our proposed models (Table 5) and the baseline models using top selected genes (Table 6, Table 7) and all genes (B_SVM_Concat, B_RF_Concat, BK_SVM_Concat, B_DNN_DBN). The models with the best results are bolded.
| Classifier | Accuracy (%) | Classifier | AUC |
|---|---|---|---|
| B_DCNN_Concat (all genes) | 96.3 | B_DCNN_Concat (all genes) | 0.993 |
| B_DCNN_Concat (top 150 genes) | 96.3 | B_DCNN_Concat (top 150 genes) | 0.992 |
| B_SVM_Concat (top 100 genes) | 95.7 | B_ SVM_Concat (top 100 genes) | 0.940 |
| B_RF_CNA (all genes) | 81.1 | B_RF_CNA (all genes) | 0.818 |
| B_RF_GENE (all genes) | 96.3 | B_RF_GENE (all genes) | 0.991 |
| B_RF_Concat (top 250 genes) | 97.5 | B_RF_Concat (top 450 genes) | 0.995 |
| B_SVM_Concat (all genes) | 90.4 | B_SVM_Concat (all genes) | 0.838 |
| BK_SVM_Concat (all genes) | 96.5 | BK _SVM_Concat (all genes) | 0.952 |
| B_RF_Concat (all genes) | 95.4 | B_RF_Concat (all genes) | 0.991 |
| B_DNN_DBN (all genes) | 86.4 | B_DNN_DBN (all genes) | 0.522 |