| G. Vonitsanos, A. Kanavos, Ph. Mylonas |
| Social Media Insights into Climate Change: Sentiment Analysis Using VADER and RoBERTa |
| 19th International Workshop on Semantic and Social Media Adaptation & Personalization (SMAP 2024), 21-22 November 2024, Athens, Greece |
|
ABSTRACT
|
| Sentiment analysis, a critical branch of Natural Language Processing (NLP), is pivotal for uncovering the emotional undertones within textual data, thereby revealing public sentiments on diverse topics. This study conducts a comparative analysis of two prominent sentiment analysis tools—VADER, a lexicon and rule-based approach, and RoBERTa, a transformerbased deep learning model. It focuses on their efficacy in analyzing tweets related to climate change, a topic of global significance that elicits a wide range of public opinions. The nuanced and dynamic nature of social media language poses unique challenges, particularly in contexts such as climate change discussions. We assess how effectively each model discerns positive, neutral, and negative sentiments across different categories of climate change-related tweets, delineated into Pro, Neutral, News, and Anti stances. Our findings indicate that RoBERTa generally outperforms VADER in capturing contextual nuances and providing detailed sentiment classifications. This detailed capability allows RoBERTa to better reflect the complex public opinions on climate change, offering this way invaluable insights to policymakers, researchers, and environmental advocates. This study not only aids in better understanding and engaging with public discourse on social media but also highlights the potential of advanced NLP tools in shaping environmental communication and policy formulation.
|
| 21 November , 2024 |
| G. Vonitsanos, A. Kanavos, Ph. Mylonas, "Social Media Insights into Climate Change: Sentiment Analysis Using VADER and RoBERTa", 19th International Workshop on Semantic and Social Media Adaptation & Personalization (SMAP 2024), 21-22 November 2024, Athens, Greece |
[  PDF] [
BibTex] [
Print] [
 Back] |
