Creation and evaluation of timelines for longitudinal user posts

Anonymous

Abstract: There is increasing interest to work with user generated content in social media and especially textual posts over time. Currently there is no consistent way of segmenting user posts into timelines in a meaningful way that can improve the quality and cost of manual annotation. Here we propose a set of methods for segmenting longitudinal user posts into timelines that are likely to contain interesting moments of change in a user's behaviour based on the content they have shared online and their online activity. We also propose a framework for evaluating the timelines returned in terms of containing candidate moments of change in close proximity to manually annotated timelines dense in such moments of change. Finally, we present a discussion of the linguistic content of highly ranked timelines.

Paper Type: long

Consent To Share Data: no