Papers for

public health social media analysts

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Metrics measure how well topics match short social media posts

Document Topic Alignment Metrics for Evaluating Topic Models of Short-Text Public Health Communications on Social Media

Abstract: Topic models are widely used to analyze public health-related social media short texts, yet their evaluation remains dominated by metrics that focus entirely on generated topics alone. There is a lack of metrics that quantitatively assess whether assigned topics meaningfully represent the corresponding short-text posts. We propose Document-Topic Alignment metrics (DoTA), an assignment-aware evaluation framework comprising metrics that measure semantic alignment between documents (posts) and their assigned topics. We also introduce margin-based and discriminative variants that capture topic assignment confidence and distinguishability. We evaluate DoTA across five topic models on three public health-related social media datasets from X and compare DoTA metrics with conventional topic-based metrics. Results show that DoTA provides complementary evaluation cues and aligns meaningfully with human evaluations. These findings establish the need for assignment-aware evaluation and demonstrate that the addition of DoTA enables a more comprehensive and practically meaningful evaluation for assessing short-text topic modeling performance.

Sun 13 SeptComputation and LanguageSocial and Information Networks
The gist
It can be hard to check if computer programs that sort short social media posts about health into topics are doing a good job. The common ways to check only look at the topics themselves, not if the posts actually fit those topics. The authors created new tools called DoTA that look at how well each post matches its assigned topic, including how confidently it was assigned. They tested these tools on health-related social media data and found that DoTA gives useful extra information that lines up well with how humans judge the matches.
Open 2609.14256v1