Table 5.

The performance of different models.

Model	Accuracy	Precision	Recall	F1-score
Linear discriminant analysis (9)	0.625	0.60	0.75	0.67
DistilBert	0.48	0.51	0.48	0.48
ERNIE (39)	0.42	0.46	0.42	0.30
DistilBert +CNN	0.58	0.34	0.58	0.43
DistilBert+RF	0.79	0.79	0.79	0.79
DistilBert+SVM	0.625	0.629	0.625	0.622
DistilBert+Ada	0.73	0.73	0.73	0.73
ERNIE+Pause (10)*	0.896	0.952	0.833	0.889
DistilBert+ LR	0.88	0.88	0.88	0.87

^*ERNIE+Pause (10) is the model of a champion, distilBert +LR is our method, RF and Ada are the abbreviations of random forest and adaboost classifier, respectively.

The best performance in a column of measure.