Lightweight Twitter Emotion Analysis Using DistilBERT for Social Media Health Alerts
Year: 2026
Type: conference-paper
Abstract: With the growth of social media and its global outreach a large volume of data is exchanged over the internet. These ideas could be good, bad and in some cases ugly. Prolonged exposure to negative ideas/comments/posts sometimes shapes an individual’s mental wellbeing. Teenagers, age group of people who are not so emotionally mature; when allowed access to social media they often times gets affected by its worst side. Political propaganda1, Body shape issues2& cyber bullying can change the trajectory of their life, affect self-confidence and even spawn self-harm thoughts. One way to avoid them is to have a much aware consumption. This project ‘Mood-Mirror’ aims to provide a tool using which users can know the emotional health of their social media feed and be much more aware of the content they consume. This tool will nudge users with a pop-up when it senses high negative flags over a long period of time. In this paper by combining the power of existing powerful NLP tools we will be developing a model that will power Mood-Mirror’s AI capabilities of categorizing posts/comments into different emotional classes.
Institutions: Galgotias University, ABES Engineering College
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7203785668