Table 4. Open in new tab Overview of...

	Accuracy	Precision	Recall	F1-score	AUROC
Neutral vs. others
Random forest	0.652	0.651	0.652	0.651	0.718
Gradient boosting	0.614	0.642	0.614	0.616	0.679
XGBoost	0.584	0.603	0.584	0.587	0.640
Logistic regression	0.622	0.633	0.622	0.625	0.643

Negative vs. positive
Random forest	0.681	0.788	0.681	0.712	0.751
Gradient boosting	0.664	0.789	0.664	0.699	0.738
XGBoost	0.659	0.795	0.661	0.696	0.721
Logistic regression	0.664	0.766	0.664	0.697	0.657

	Accuracy	Precision	Recall	F1-score	AUROC
Neutral vs. others
Random forest	0.652	0.651	0.652	0.651	0.718
Gradient boosting	0.614	0.642	0.614	0.616	0.679
XGBoost	0.584	0.603	0.584	0.587	0.640
Logistic regression	0.622	0.633	0.622	0.625	0.643

Negative vs. positive
Random forest	0.681	0.788	0.681	0.712	0.751
Gradient boosting	0.664	0.789	0.664	0.699	0.738
XGBoost	0.659	0.795	0.661	0.696	0.721
Logistic regression	0.664	0.766	0.664	0.697	0.657

Table 4.

Overview of hub-based sentiment prediction performance.