Automated Measures of Sentiment via Transformer- and Lexicon-Based
Sentiment Analysis (TLSA)
Abstract
The last decade witnessed the proliferation of automated content
analysis in communication research. However, existing computational
tools have been taken up unevenly, with powerful deep learning
algorithms such as transformers rarely applied as compared to
lexicon-based dictionaries. To enable communication scholars to adopt
modern computational methods for valid and reliable sentiment analysis,
we propose a free and open web service named transformer- and
lexicon-based sentiment analysis (TLSA). TLSA integrates different
sentiment analysis tools including transformer-based deep learning
models and sentiment dictionaries. TLSA enables users with limited
computational resources to use these tools and validate their results at
once. Two cases showed that transformer-based models provided more
accurate measurement of sentiments in tweets, compared to lexicon-based
methods. The performance of different tools varied to a large extent
based on the dataset, suggesting the importance to validate different
sentiment tools in a specific study. TLSA is expected to empower
communication scholars to reap the benefit of state-of-the-art
computational methods for valid measurement and insight discovery.