Skip to main content

Table 7 Key hyperparameters of the six traditional classifiers and their optimal value after grid search

From: MAGCNSE: predicting lncRNA-disease associations using multi-view attention graph convolutional network and stacking ensemble model

Method Optimal hyperparameters
RandonForest max_feature=10; min_sample_split=2; n_estimators=2000
ExtraTrees max_feature=10; min_sample_split=2; n_estimators=2000
XGBoost learning_rate=0.05; max_depth=4; gamma=0; n_estimators=1000
LightGBM learning_rate=0.15; max_depth=10; num_leaves=31; n_estimators=200
CatBoost depth=3; iteration=800; learning_rate=0.1; border_count=32; l2_leaf_reg=5
LogisticRegression C=20.0; max_iter=40; penalty=‘l2’