Skip to main content

Table 2 Hyper-parameters of baseline methods

From: Disease risk analysis for schizophrenia patients by an automatic AHP framework

Methods Hyper-parameters
Random Forest num of trees: 1000, num of attr consider at each split: 5
Neural Network Neurons of hidden layers: 100, activation: Relu, solver: Adam, regularization, learning rate: 0.001, iters: 200
Logistic Regression regularization type: ridge(L2), strength: C = 1
SGD Loss function: logistic regression, regularization method: Elastic Net, \(\epsilon\): 0.1, iters:1000
kNN K: 9, metric: Euclidean, weight: Uniform
SVM RBF, Kernel:\(exp(-g|x-y{|}^{2})\), C: 1.00,: 0.1, iteration limit: 100