Characterizing and curating conversation threads: Expansion, focus, volume, re-entry

摘要: Discussion threads form a central part of the experience on many Web sites, including social networking sites such as Facebook and Google Plus knowledge creation Wikipedia. To help users manage challenge allocating their attention among discussions that are relevant to them, there has been growing need for algorithmic curation on-line conversations --- development automated methods select subset present user. Here we consider two key sub-problems inherent in conversational curation: length prediction predicting number comments discussion thread will receive novel task re-entry whether user who participated later contribute another comment it. The first these arises estimating how interesting is, sense generating lot conversation; second can determine should be kept notified progress which they have already contributed. We develop evaluate range approaches tasks, based an analysis network structure arrival pattern participants, well dichotomy long threads. find both learning-based using sources information yield improvements all performance metrics used.