In social media, demographic inference is a critical task in order to gain a better understanding of a cohort and to facilitate interacting with one's audience. Most previous work has made independence assumptions over topological, textual and label information on social networks. In this work, we employ recursive neural networks to break down these independence assumptions to obtain inference about demographic characteristics on Twitter. We show that our model performs better than existing models including the state-of-theart.
This paper describes the Data61-CSIRO text classification systems submitted as part of the CLPsych 2016 shared task. The aim of the shared task is to develop automated systems that can help mental health professionals with the process of triaging posts with ideations of depression and/or self-harm. We structured our participation in the CLPsych 2016 shared task in order to focus on different facets of modelling online forum discussions: (i) vector space representations; (ii) different text granularities; and (iii) fine-versus coarse-grained labels indicating concern. We achieved an F1score of 0.42 using an ensemble classification approach that predicts fine-grained labels of concern. This was the best score obtained by any submitted system in the 2016 shared task. * This work was performed while Yufei was at CSIRO.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.