Table 4 Outcomes of implementing different feature selection and classification methods on the combine features – CF.
Model | Accuracy | Val Accuracy | Precision | Val Precision | Recall | Val Recall | Loss | Val Loss |
|---|---|---|---|---|---|---|---|---|
(A) LSTM Classifier | ||||||||
DenseNet201 | 84.29 | 86.30 | 81.52 | 82.17 | 100.0 | 100.0 | 0.42 | 0.39 |
Inception ResNet V2 | 77.12 | 78.70 | 71.32 | 71.74 | 100.0 | 100.0 | 0.62 | 0.59 |
MobileNet V2 | 71.81 | 71.09 | 63.63 | 62.61 | 100.0 | 100.0 | 0.75 | 0.78 |
ResNet 152 V2 | 78.39 | 76.74 | 72.16 | 70.22 | 99.16 | 97.30 | 0.56 | 0.61 |
Combined Features | 91.23 | 89.91 | 89.86 | 86.52 | 100.0 | 100.0 | 0.29 | 0.37 |
Model | Accuracy | Val Accuracy | Precision | Val Precision | Recall | Val Recall | Loss | Val Loss |
|---|---|---|---|---|---|---|---|---|
(B) Bi-LSTM Classifier | ||||||||
DenseNet201 | 84.39 | 86.30 | 81.55 | 82.61 | 99.5 | 100.0 | 0.41 | 0.38 |
Inception ResNet V2 | 77.54 | 79.57 | 71.55 | 71.96 | 100.0 | 100.0 | 0.61 | 0.58 |
MobileNet V2 | 71.55 | 70.65 | 63.10 | 62.17 | 100.0 | 100.0 | 0.73 | 0.76 |
ResNet 152 V2 | 78.91 | 77.39 | 72.98 | 71.09 | 97.93 | 97.62 | 0.55 | 0.60 |
Combined Features | 91.27 | 89.13 | 89.93 | 86.96 | 100.0 | 98.67 | 0.28 | 0.35 |
Model | Accuracy | Val Accuracy | Precision | Val Precision | Recall | Val Recall | Loss | Val Loss |
|---|---|---|---|---|---|---|---|---|
(C) DenseNet Classifier | ||||||||
DenseNet201 | 84.68 | 86.30 | 82.17 | 83.91 | 100.0 | 100.0 | 0.40 | 0.36 |
Inception ResNet V2 | 78.36 | 79.13 | 73.50 | 73.48 | 99.19 | 99.20 | 0.59 | 0.58 |
MobileNet V2 | 71.68 | 72.17 | 65.03 | 65.00 | 99.49 | 98.72 | 0.73 | 0.76 |
ResNet 152 V2 | 78.78 | 75.44 | 73.50 | 71.52 | 98.66 | 97.79 | 0.55 | 0.60 |
Combined Features | 91.10 | 88.04 | 89.99 | 86.74 | 98.83 | 98.45 | 0.28 | 0.35 |
