You Too?! Mixed-Initiative LDA Story Matching to Help Teens in Distress

Abstract

Adolescent cyber-bullying on social networks is a phenomenon that has received widespread attention. Recent work by sociologists has examined this phenomenon under the larger context of teenage drama and it's manifestations on social networks. Tackling cyber-bullying involves two key components – automatic detection of possible cases, and interaction strategies that encourage reflection and emotional support. Key is showing distressed teenagers that they are not alone in their plight. Conventional topic spotting and document classification into labels like "dating" or "sports" are not enough to effectively match stories for this task. In this work, we examine a corpus of 5500 stories from distressed teenagers from a major youth social network. We combine Latent Dirichlet Allocation and human interpretation of its output using principles from sociolinguistics to extract high-level themes in the stories and use them to match new stories to similar ones. A user evaluation of the story matching shows that theme-based retrieval does a better job of finding relevant and effective stories for this application than conventional approaches.

Modeling Empathic Similarity in Personal Narratives

“…Document similarity is a well-defined task in NLP (Salton et al, 1997;Damashek, 1995;Deerwester et al, 1990;Landauer and Dumais, 1997), but few have applied this work to matching personal narratives based on shared emotional experiences (Chaturvedi et al, 2018;Lin et al, 2014). One study used Latent Dirichlet Allocation (LDA) to cluster cyberbullying stories and match these stories based on similarity in theme (Dinakar et al, 2012), but discovered that only 58.3% found the matched story to be helpful if provided to the narrator of the original story.…”

“I Will Not Drink With You Today”: A Topic-Guided Thematic Analysis of Addiction Recovery on Reddit

“…Researchers have employed the use of computational methods to derive a variety of topics from large data corpi for some time (e.g. [8,30,31,34,68,74]). LDA has been particularly useful to HCI researchers in identifying latent topics in Reddit communities (e.g., [2,3,82,83]), and has been described by Ammari et al [3] as part of a 'roadmap' for using computational techniques to better understand social relationships online.…”