Table 3 Model Parameters.

From: Adversarial susceptibility analysis for water quality prediction models

Model

Key Hyperparameters

Random Forest

n_estimators = 500, max_depth = 10, min_samples_split = 2, random_state = 42

XGBoost

n_estimators = 300, max_depth = 6, learning_rate = 0.1, subsample = 0.8, gamma = 0

MLP

hidden_layer_sizes=(100, 50), activation=’relu’, solver=’adam’, max_iter = 1000

TabNet

n_d = 32, n_a = 32, n_steps = 5, gamma = 1.5, lambda_sparse = 1e-3, max_epochs = 100