Table 8 Final parameters of the two types of machine learning classifiers used
From: High-dimensional detection of imaging response to treatment in multiple sclerosis
Classifier | Final parameters and settings |
|---|---|
SVM | Gaussian radial basis function kernel, penalty term C = 10 (with balanced class weighting), kernel coefficient γ = number of features−1 |
ERT | Gini impurity as the tree-splitting metric, number of trees = 100, number of features to consider when looking for the best split Mf = number of features1/2, balanced class weighting |
