Table 3. The key parameter settings for different methods.

Method	Parameters
SVM (Depuru, Wang & Devabhaktuni, 2011)	Kernel type: RBF kernel; regularization parameter C = 1; Kernel coefficient γ = 0.92.
Wide & Deep CNN (Zheng et al., 2017)	The size of the wide component α = 60; the size of the CNN component β = 120; the number of filters γ = 15.
LSTM (Munawar et al., 2021)	Embedding dimension de = 256; the number of hidden units nh = 128.
CNN-LSTM (Almazroi & Ayub, 2021)	Embedding dimension de = 256; the number of hidden units nh = 128.
Autoencoder (Takiddin et al., 2022)	Encoder Layers ne = 4; decoder layers nd = 2.
Transformer (Ours)	Key dimension dk = 256; number of heads h = 4.