Fig. 9
From: Large-scale comparison of machine learning methods for profiling prediction of kinase inhibitors
Comparison of the prediction results between fusion models and single models. The fusion models are constructed based on RF (A), KNN (B), and NB (C) models using voting and stacking methods