User’s Emotion Profiling in Web Browsing Behavior

Abstract

AbstractThis study presents a method of user's emotion profiling in Web browsing behavior, based on our assumption that a user can be continuously affected by digital information through the Web site access. There are many studies on the user’s emotion model, however, it is still an important issue to quantitatively calculate the amount of emotion change caused by accessing the digital information. This study focuses on designing a user emotion model where a state of user’s emotion changes according to contents that the user is accessing on the Web. When a user’s emotion model is too sensitive, the user’s emotion state is frequently updated, every time the user accesses the Web contents. To avoid this problem, our proposed method calculates the current value of user’s emotion based on the gap between an emotional polarity of the Web contents that a user is accessing and the user’s emotion profiled in advance for the corresponding topic. By this function of gap calculation, the user’s emotion model can be robust for the excessive frequency of user’s emotion changes in the context of user’s emotion profiling.