Skip to main content
. 2022 Aug 29;129:109588. doi: 10.1016/j.asoc.2022.109588

Table 5.

A summary of the proposed models and training hyperparameters with best results.

Dataset Models Loss function Activation function Classifier Optimizer (lr:learning rate) Batch Epochs
Dataset 1 WCL + CheXNet WCL RELU Softmax Adam (lr = 104) 32 50
SMOTE + DenseNet201 Categorical cross-entropy RELU Softmax Adam (lr = 104) 8 72

Dataset 2 WCL + DenseNet201 WCL RELU Softmax Adam (lr = 104) 8 20
SMOTE + DenseNet201 Categorical cross-entropy RELU Softmax Adam (lr = 104) 8 50

Dataset 3 WCL + VGG19 WCL RELU Softmax Adam (lr = 104) 8 20
SMOTE + VGG19 Categorical cross-entropy RELU Softmax Adam (lr = 104) 8 50