Table 1.
Performance of lncRNA_Mdeep and other model architectures in the 10CV test.
| ACC (%) | Sn (%) | Sp (%) | MCC | |
|---|---|---|---|---|
| OFH_DNN | 95.74 ± 1.70 | 94.44 ± 4.89 | 97.04 ± 2.15 | 0.9171 ± 0.0307 |
| k-mer_DNN | 96.53 ± 0.41 | 96.40 ± 1.11 | 96.66 ± 0.78 | 0.9307 ± 0.0082 |
| One-hot_CNN | 95.82 ± 0.33 | 97.01 ± 0.96 | 94.63 ± 1.19 | 0.9169 ± 0.0064 |
| OFH_DNN + k-mer_DNN | 95.97 ± 2.49 | 96.87 ± 1.05 | 95.06 ± 5.71 | 0.9211 ± 0.0449 |
| k-mer_DNN + One-hot_CNN | 98.36 ± 0.16 | 98.70 ± 0.42 | 98.03 ± 0.50 | 0.9674 ± 0.0033 |
| OFH_DNN + One-hot_CNN | 97.60 ± 1.26 | 97.78 ± 1.58 | 97.43 ± 2.33 | 0.9526 ± 0.0248 |
| Decision fusion | 98.42 ± 1.12 | 99.24 ± 0.45 | 97.60 ± 2.59 | 0.9689 ± 0.0212 |
| lncRNA_Mdeep | 98.73 ± 0.41 | 98.95 ± 0.54 | 98.52 ± 0.92 | 0.9748 ± 0.0080 |
OFH_DNN, the DNN model with the OFH feature as input. k-mer_DNN: the DNN model with the k-mer feature as input. One-hot_CNN: the CNN model with one-hot encoding as input. ACC, accuracy. Sn, sensitivity. Sp, specificity. MCC, Matthew’s correlation coefficient.