Table 2 The optimized hyperparameters used for DL and ML techniques
From: Novel transfer learning based bone fracture detection using radiographic images
Model | Hyperparameter values |
|---|---|
CNN | activation=’sigmoid’, optimizer=’adam’, loss=’binary_crossentropy’ |
MobLG-Net | weights = ’imagenet’, activation=’softmax’, |
LGB | n_estimators=10, boosting_type=’gbdt’, max_depth=−1, num_leaves=31 |
KNN | n_neighbors=2, weights=’uniformn’, metric=’minkowski’, leaf_size=30, p=2 |
LR | penalty=’l2’, max_iter=100, solver=’lbfgs’, fit_intercept=True |
RF | max_depth=300, criterion=’gini’, min_samples_split=2, n_estimators=300, random_state=0 |