Towards Accurate Deceptive Opinions Detection Based on Word Order-Preserving CNN

Table 3

Hyper-parameter setting.

Hyper-parameter	Description	Value
d	word vector dimension	100
	convolution width	3,4,5
H	number of convolutions	64,64,64
L2	L2 regularization lambda	0.5
dropout	dropout keep probability	0.5
mini-batch	mini-batch gradient descent	50