Table 1.

Best hyperparameters for the classifiers

Classifier	Hyperparameters
Deep learning classifiers
LSTM_onFly	Optimizer=Adam, batch size=64, dropout rate=0, word embedding=trained on the fly, recurrent layer=single directional LSTM
LSTM_Pre	Optimizer=Adam, batch size=64, dropout rate=0, word embedding=pretrained on the whole corpus, recurrent layer=single directional LSTM
LSTM_Bi	Optimizer=Adam, batch size=64, dropout rate=0, word embedding=pretrained on the whole corpus, recurrent layer=bidirectional LSTM
GRU	Optimizer=Adam, batch size=64, dropout rate=0, word embedding=pretrained on the whole corpus, recurrent layer=single directional LSTM
Conventional classifiers
KNN	Number of neighbors=7
LR	L2 penalty parameter=10
NB	Smoothing parameter alpha=0
RF	Maximum depth of a tree=6 Minimum number of samples required to split an internal node=5 Minimum number of samples required to be at a leaf node=5
SVC	Kernel=linearL 2 penalty parameter=30

GRU: gated recurrent unit; KNN: K-nearest Neighbor; LR: logistic regression; LSTM: long short-term memory; NB: Naive Bayes; RF: random forest; SVC: Support Vector Machine for classification.