Skip to main content

Table 6 Proposed model best parameters

From: Deep learning for emotion analysis in Arabic tweets

Dropout rates 0.5,0.5,0.5
Optimization Adam
Loss “mse”
Number of BiLSTM layers 3
Activation Function in BiLSTMS Relu
Activation Function in output Layer Sigmoid
Epochs 50
Batch Size 200
Learning Rate 0.001