Table 8. Performances of sentiment analysis models on NLPCC dataset.

Method	Performance (%)
	Precision	Recall	F1-score
NB	63.82	61.90	58.84
LR	57.80	56.99	56.53
SVM	63.86	63.49	59.85
TextCNN	68.50	67.37	67.62
LSTM	69.48	72.29	70.48
SVMRDEV	69.05	75.03	70.16
TextCNNRDEV	70.36	73.20	71.10
LSTMRDEV	71.10	75.60	72.39