A Multilayer Naïve Bayes Model for Analyzing User’s Retweeting Sentiment Tendency

Table 2

Information gains of features.

Types of features	Features	Information gains
Profile-based	#bifollowers	0.254
	#followers	0.231
	#followees	0.257
	#posts	0.201
	Province	0.105
	City	0.110
	Gender	0.148
	Created time of user’s account	0.092
	Verified type of user’s account	0.102
Relation-based	Dynamic Salton metrics	0.485
	Dynamic interaction frequency	0.503
Emotion-based	#positive emotional words	0.647
	#negative emotional words	0.622
	Recent mood statistics	0.694
	Emotion divergence	0.723