Sentimental Analysis of Industry 4.0 Perspectives Using a Graph-Based Bi-LSTM CNN Model

Table 1

The hyperparameter feature settings.

Sl no.	Layers	Parameters
1	Embedding	input_dim = 5000, output_dim = 20,
		input_length = 30
2	SimpleRNN	Neuron units = 100
3	LSTM	Neuron units = 100
4	CNN	nb_filter = 20, filter_length = 3,
		activation = ‘relu’
5	Max-pooling	pool_length = 2
6	Dropout(Layer 1)	Units: 0.5
7	Dropout(Layer 2)	Units: 0.3
8	Dense(Layer 1)	units = 20, activation = ‘relu’
9	Dense(Layer 2)	units = 1, activation = ‘sigmoid’
10	model.compile()	loss = ‘binary_crossentropy’, optimizer = ‘Adam’
11	model.fit()	batch_size = 32, epochs = 8
12	model.predict()	batch_size = 32, verbose = 1
13	model.evaluate()	verbose = 1