Table 4 Parameter grids and selected optimal hyperparameters obtained from grid search.

From: Machine learning-based models for screening of anemia and leukemia using features of complete blood count reports

Algorithm

Parameter Grid

Output

DT

param_grid = {‘criterion’: [‘gini’, ‘entropy’],

‘max_depth’: [None, 5, 10, 20],

‘min_samples_leaf’: [1, 2, 4],

‘min_samples_split’: [2, 5, 10]}

{‘criterion’: ‘gini’,

‘max_depth’: 10, ‘min_samples_leaf’: 1, ‘min_samples_split’: 2}

RF

param_grid = {‘n_estimators’: [50, 100, 200],

‘criterion’: [‘gini’, ‘entropy’],

‘max_depth’: [None, 5, 10],

‘min_samples_leaf’: [1, 2],

‘min_samples_split’: [2, 5]}

{‘n_estimators’: 100,

‘criterion’: ‘gini’,

‘max_depth’: ‘None’, ‘min_samples_leaf’: 1, ‘min_samples_split’: 2}

GBM

param_grid = {‘n_estimators’: [50, 100, 200],

‘criterion’: [‘friedman_mse’, ‘squared_error’],

‘learning_rate’: [0.01, 0.1, 0.2],

‘max_depth’: [3, 5],

‘min_samples_leaf’: [1, 2],

‘min_samples_split’: [2, 5]}

{‘n_estimators’: 100,

‘criterion’: ‘friedman_mse’, ‘learning_rate’: 0.01, ‘max_depth’: 3, ‘min_samples_leaf’: 1, ‘min_samples_split’: 2}

SVM

param_grid = {‘C’: [1, 10, 20, 30],

‘kernel’: [‘rbf’, ‘poly’],

‘gamma’: [‘scale’, ‘auto’]}

{‘C’: 30,

‘kernel’: ‘rbf’,

‘gamma’: ‘scale’}

LR

param_grid = {‘C’: [0.01, 0.1, 1, 10],

‘penalty’: ‘l1’, ‘l2’, ‘elasticnet’,

‘solver’: [‘lbfgs’, ‘liblinear’]}

{‘C’: 0.01,

‘penalty’: ‘l2’,

‘solver’: ‘lbfgs’}

MLP

param_grid = {‘hidden_layer_sizes’: [(200, 100, 50), (100, 50, 10)],

‘activation’: [‘relu’, ‘tanh’],

‘alpha’: [0.0001, 0.001, 0.01]

‘max_iter’: [200, 500, 1000]}

{‘hidden_layer_sizes’: (100, 50, 10),

‘activation’: ‘relu’,

‘alpha’: 0.01

‘max_iter’: 1000}

Back to article page