Skip to main content

Table 1 The choice of hyperparameters of each model

From: Using machine learning models to improve stroke risk level classification methods of China national stroke screening

Machine learning modelsHyperparametersValues to be selectedOptimum Value
Decision tree (C4.5)confidence factor used for pruning (C); minimum number of instances of each leaf (N)C = 0.1,0.15,0.2,0.25, 0.3; N = 2,3,4,5,6C = 0.25; N = 2
Neural networkthe size of network (number of hidden nodes, H); gradient descent (D).H = 3, 4, 8, 10, 20, 50, 100 and D = 0.00001, 0.001, 0.1, 0.5, 0.9H = 4; D = 0.1
Random forestthe depth of the tree(T); number of tree models(N)T = 1, 2, 3, 5, 10; N = 100, 200, 300, 500T = 8; N = 300
Bagging with C4.5 decision treethe sampling ratio (P); number of sub-classifiers(N)P = 70, 80, 90, 95, 100%; N = 100, 150, 200, 300, 500P = 90%; N = 200
Boosting with C4.5 decision treethe number of sub-classifiers(N)N = 10, 30, 50, 100N = 30