Table 4 Detailed outcomes of the four methods mentioned above with the input of matrices of structural profile properties
From: A successful hybrid deep learning model aiming at promoter identification
Organism | Method | Sn (%) | Sp (%) | Acc (%) | Mcc |
|---|---|---|---|---|---|
Human | DSPN | 80.60 | 90.36 | 86.73 | 0.7144 |
CNNs | 82.63 | 86.76 | 85.22 | 0.6874 | |
ResNet | 80.81 | 86.84 | 84.59 | 0.6725 | |
GoogLeNet | 78.31 | 89.56 | 85.37 | 0.6847 | |
Plants | DSPN | 75.24 | 90.33 | 84.41 | 0.6695 |
CNNs | 81.76 | 85.50 | 83.23 | 0.6607 | |
ResNet | 71.37 | 90.80 | 83.18 | 0.6428 | |
GoogLeNet | 67.51 | 93.69 | 83.42 | 0.6500 |