Skip to main content

Table 2 Optimization of the machine learning classifiers

From: Distinguishing crystallographic from biological interfaces in protein complexes: role of intermolecular contacts and energetics for classification

  Classification accuracy on the MANY dataset Classification accuracy on the independent DC dataset
Bagging 0.92 0.73
Random Forest 0.92 0.74
Adaptive Boosting 0.92 0.74
Gradient Boosting 0.93 0.74
Neural Network 0.91 0.75
  1. The maximum accuracy reached by optimizing the settings is reported by each classifier for the Many (as average over the 10-fold cross-validation) and the DC datasets