Table 9. Predictive performance of different approaches based on the variance-stability transformation and the binary transformation for the gut microbiome data related to PD.
Dataset | Variance-stability transformation | Binary transformation | ||||
---|---|---|---|---|---|---|
Method | ER | AUC | AUPRC | ER | AUC | AUPRC |
LASSO | 0.248 | 0.811 | 0.872 | 0.221 | 0.812 | 0.856 |
ZINB+GLM | 0.291 | 0.720 | 0.771 | 0.318 | 0.706 | 0.758 |
TPNB+GLM | 0.352 | 0.663 | 0.725 | 0.318 | 0.703 | 0.754 |
NB+GLM | 0.268 | 0.743 | 0.765 | 0.264 | 0.758 | 0.768 |
ZINB+LASSO | 0.249 | 0.801 | 0.861 | 0.199 | 0.872 | 0.912 |
TPNB+LASSO | 0.256 | 0.810 | 0.873 | 0.218 | 0.853 | 0.890 |
NB+LASSO | 0.222 | 0.810 | 0.871 | 0.249 | 0.787 | 0.818 |