COVID-19 vaccine hesitancy: a social media analysis using deep learning

Area of Science:

• Public Health
• Computational Linguistics
• Epidemiology

Background:

• Vaccine hesitancy is a major global health concern, exacerbated by misinformation on social media.
• Social media platforms significantly influence public vaccine decisions, posing challenges for health campaigns.

Purpose of the Study:

• To evaluate machine learning and deep learning models for identifying vaccine-hesitant tweets during the COVID-19 pandemic.
• To compare the effectiveness of different computational methods in detecting vaccine hesitancy online.

Main Methods:

• Utilized Long Short-Term Memory (LSTM) and Recurrent Neural Network (RNN) models.
• Compared deep learning approaches against traditional machine learning models for text classification.
• Analyzed tweets related to vaccination during the COVID-19 pandemic.

Main Results:

• Deep learning models, specifically LSTM and RNN, achieved higher accuracy in identifying vaccine-hesitant tweets.
• LSTM and RNN models demonstrated an 86% accuracy rate.
• Traditional machine learning models achieved an 83% accuracy rate.

Conclusions:

• LSTM and RNN models are more effective than traditional machine learning for detecting vaccine hesitancy in social media data.
• These findings can inform public health strategies to combat vaccine misinformation online.