Sentiment Prediction of Textual Data Using Hybrid ConvBidirectional-LSTM Model

Table 3

Hyperparameter setting.

Parameters	Values
Embedding dimensions	GloVe = 300
Kernel size	5
Filter	128
Activation	Swish
Pool size	3
Dropout	0.3
Bi-LSTM output size	512
Batch size	128
Epoch	15
Optimizer	Adam
Learning rate	0.001
Loss function	Categorical_crossentropy