Table 3 Performance comparison of 188-D features and 473-D features.
From: Resistance gene identification from Larimichthys crocea with machine learning techniques
Feature extraction method | Dimension | Training set | Accuracy | ||||
|---|---|---|---|---|---|---|---|
Resistance gene | Non-Resistance gene | SN | SP | Mcc | Accuracy (%) | ||
188-D | 188 | 6720 | 6720 | 0.831 | 0.687 | 0.523 | 75.878 |
Pse-AAC | 30 | 6720 | 6720 | 0.761 | 0.627 | 0.392 | 69.4345 |
473-D | 473 | 178 | 226 | 0.371 | 0.752 | 0.133 | 58.4158 |
Feature extraction method | Dimension | test set | |||||
Resistance gene | Non-Resistance gene | Accuracy (%) | |||||
188-D | 188 | 0 | 3308 | 69.347 | |||
Pse-AAC | 30 | 0 | 3308 | 60.9129 | |||
473-D | 473 | 20 | 20 | 55.0 | |||
