Characterizing high-impact features for content retention in social web applications

Kaweh Djafari Naini, Ricardo Kawase, Nattiya Kanhabua, Claudia Niederée

Abstract: One of the core challenges of automatically creating Social Web summaries is to decide which posts to remember, i.e., to consider for summary inclusion and which to forget. Keeping everything would overwhelm the user and would also neglect the often intentionally ephemeral nature of Social Web posts. In this paper, we analyze high-impact features that characterize memorable posts as a first step for this selection process. Our work is based on a user evaluation for discovering human expectations towards content retention.