Multimodal Blog Sentiment Classification Based on MD-HFCE

Table 6

Comparison of dictionary experiment results.

	Precision			Recall			F1
	NTUSD	HowNet	Ours	NTUSD	HowNet	Ours	NTUSD	HowNet	Ours
Positive blog	0.711	0.767	0.857	0.747	0.76	0.812	0.729	0.764	0.834
Neutral blog	0.694	0.667	0.782	0.548	0.631	0.768	0.612	0.649	0.775
Negative blog	0.799	0.752	0.827	0.73	0.757	0.791	0.763	0.754	0.801

Bold values intuitively compare and emphasize the effectiveness of the fuzzy dictionary established in this article.