Table 8 ML and DL multi-class classification results for SARS-CoV-2 proteins using the real experimental dataset and the GAN augmented dataset, respectively.
From: DeepATsers: a deep learning framework for one-pot SERS biosensor to detect SARS-CoV-2 virus
Model | Accuracy | Precision | Sensitivity | Specificity | f1-score | |||||
|---|---|---|---|---|---|---|---|---|---|---|
Real data | Augm. data | Real data | Augm. data | Real data | Augm. data | Real data | Augm. data | Real data | Augm. data | |
RF | 0.5600 | 0.9437 | 0.4933 | 0.9465 | 0.5600 | 0.9437 | 0.7904 | 0.9851 | 0.5040 | 0.9432 |
GBM | 0.5200 | 0.9500 | 0.3404 | 0.9551 | 0.5200 | 0.9500 | 0.7904 | 0.9872 | 0.4095 | 0.9505 |
SVM | 0.6000 | 0.9375 | 0.4745 | 0.9409 | 0.6000 | 0.9375 | 0.8281 | 0.9838 | 0.5268 | 0.9380 |
KNN | 0.6400 | 0.9187 | 0.5806 | 0.9245 | 0.6400 | 0.9187 | 0.8520 | 0.9781 | 0.6038 | 0.9177 |
DeepATsers | 0.6000 | 0.9750 | 0.5529 | 0.9757 | 0.6000 | 0.9750 | 0.7594 | 0.9984 | 0.5479 | 0.9750 |