MIT Press

Skip to Main Content

Table 4:

Evaluation of tweet classification results in accuracy (Acc) and average F1 (Avg F1). Representations learned by various models serve as the classification features. For our model, both the topic and discourse representations are fed into the classifier.

Models	TREC		TWT16
	Acc	Avg F1	Acc	Avg F1
Baselines
BoW	0.120	0.026	0.132	0.030
TF-IDF	0.116	0.024	0.153	0.041
LDA	0.128	0.041	0.146	0.046
BTM	0.123	0.035	0.167	0.054
LF-DMM	0.158	0.072	0.162	0.052
NTM	0.138	0.042	0.186	0.068
Our model	0.259	0.180	0.341	0.269

Models	TREC		TWT16
	Acc	Avg F1	Acc	Avg F1
Baselines
BoW	0.120	0.026	0.132	0.030
TF-IDF	0.116	0.024	0.153	0.041
LDA	0.128	0.041	0.146	0.046
BTM	0.123	0.035	0.167	0.054
LF-DMM	0.158	0.072	0.162	0.052
NTM	0.138	0.042	0.186	0.068
Our model	0.259	0.180	0.341	0.269