Spoken Language Identification Using Deep Learning

Table 6

Parameter setting.

Proposed technique/parameters	Value
Convolutional neural network
Number of epochs	60
Batch size	32
Activation	ReLU
Optimizer	Adam
Dropouts	Yes
Loss	Categorical cross-entropy
Activation output	Softmax
Word embedding
Number of epochs	25
CNN model	Word embedding Keras
Loss	Categorical cross-entropy
Optimizer	Adam
Activation output	Softmax
Naïve Bayes
Type	Bernoulli Naïve Bayes
Kernel function	Bernoulli