From: Screening ovarian cancer by using risk factors: machine learning assists
Algorithm | Hyperparameter |
---|---|
ANN | Number of hidden layers: 8; learning rate: 0.8; training epoch: 100; validation threshold: 50; nominal to binary filter: true |
KNN | 3 ≤ K ≤ 7; Nearest neighbor search algorithm: Euclidean; Distance weighting: 1/distance |
J-48 | Binary split: false; number of objects: 2; confidence factor: 0.3; reduced pruning: true; number of folds: 3; Use Laplace: true |
RF | Max_Depth: 8; number of iterations: 100; calculate out of bag: true; number of randomly chosen features: 6; classifiers: decision stump |
SVM | Kernel type: RBF; calibrator: logistic; Epsilon: 1.0E−12; c:10; tolerance parameter: Num folds: − 1; RBF-gamma: 0.1 |
XG-Boost | Booster: gb-tree; nthread: MAX; eta: 0.5; Gamma: 1; max_depth: 8; mi_child_weight: 1; max delta step: 0; sub_sample:1; Lambda:1; alpha: 0; scale_pos_weight: 1; objective: binary:logistic |