Table 3 Parameters tuning of the blending ensemble model
From: m1A-Ensem: accurate identification of 1-methyladenosine sites through ensemble models
Base models | ANN | KNN | SVM | DT |
|---|---|---|---|---|
Hyper-Parameters value(s) | Hidden_layer_sizes = 5,2 Random_state = 1 Activation = relu Solver = lbfgs Learning rate = adaptive Alpha = 0.0001 | k = 3 | C = 10 Gamma = 0.0001 Kernel = rbf Coefficient = 0.0 Probability = ‘True’ Verbose = ‘False’ Random_state = none | Splitter = ‘random’ Max_depth = 80 min_samples_leaf = 4 random_state = None |
Meta classifier & its Hyper-parameter value(s) | Gradient Boost classifier n_estimators = 100, criterion = ‘mse’ | |||