Behavior-Interior-Aware User Preference Analysis Based on Social Networks

Table 2

Topic-oriented Behavior Interior Dimensions.

Dimensions	Subdimensions	Definition in our study	Measurement in our study
Motivation	Practical value [46, 47]	Degree to which the content provides useful information	Lack of information
	Entertaining value	Degree to which the content is interesting or surprising or awe-inspiring
Affect	Sentiment [48]	Valence- positive or negative
	Controversy [14]	Sentiment variation approximated opinions pro or against
Content richness	Content volume [49]	Word count in a topic	,; is word count at time ,; , is the Levenshtein distance between two tweets
	Content diversity	Nonredundancy of topic words
Virality	Hotness [14]	Transmission times and reach of impact	, and are communication count and coverage of people of a topic at time
	Trend momentum [14]	Absolute hotness difference in successive time windows

: all the tweets are the same; : none of the tweets have any content that occurred in other tweets.