Multiclass Event Classification from Text

Table 9

RNN’s hyperparameters.

Parameter	Value
RNN (LSTM) (80.3%)
Max_words	50000
Batch size	64
Embedding_dim	100
Activation function	SoftMax
Recurrent dropout	0.2
Training/testing	90%–10%
No. of epochs	05
Loss function	Sparse categorical cross-entropy
RNN (LSTM) (81%)
Max_words	30000
Batch size	128
Embedding_dim	100
Activation function	SoftMax
Recurrent dropout	0.2
Training/testing	80%–20%
No. of epochs	05
Loss function	Sparse categorical cross-entropy