Table 6 F1-Scores obtained for the best standard filtering approach and best ML classifier across the different simulations. The consensus call of at least four out of the 7 tools was the winner option for the first category while GBDT model for the second
From: NeoMutate: an ensemble machine learning framework for the prediction of somatic mutations in cancer
Method | Category | Description | S1 | S2 | S3 | S4 |
|---|---|---|---|---|---|---|
GBDT | ML | Gradient boosting decision tree | 0.9742 | 0.9762 | 0.9658 | 0.8748 |
cons_4 | Standard filtering | Consensus of > = 4 tools | 0.9139 | 0.9326 | 0.9203 | 0.8044 |