Using Deep Learning to Determine Time and Geographic Trends of Sentiments Towards Covid-19 Vaccine


JIITA, vol.8 no.4 p.1014-1021 2024, DOI: 10.22664/ISITA.2024.8.4.1014

Alex A. Diola, Robert R. Roxas
University of the Philippines Cebu

Abstract: Vaccine hesitancy is one of the challenges faced in the battle against the Covid-19 pandemic. Understanding the sentiments of the public regarding Covid-19 vaccine across various locations throughout the pandemic will allow policy makers to better craft vaccine rollout plans. This paper examines the use of deep learning models to analyze sentiments towards Covid-19 vaccine using Twitter data to analyze time and geographic trends. The LSTM model achieved 61% accuracy, the GRU model achieved 60% accuracy, and the simple RNN achieved 48% accuracy. The time graph showed that the sentiments varied in quantity but generally exhibited the same trend behavior. The geo map did not show any significant information, due to the lack of location data, for trend analysis to reliable conducted.
Keywords: Covid-19 vaccine; deep learning; geographic trend; sentiment analysis; time trend

Fullpaper:

Scroll to top