A Proposed Sentiment Analysis Deep Learning Algorithm for Analyzing COVID-19 Tweets.

Area of Science:

• Computational Social Science
• Public Health Informatics
• Natural Language Processing

Background:

• The COVID-19 pandemic caused widespread public anxiety and pressure on global resources.
• Social media platforms, like Twitter, became dominant channels for information dissemination, influencing public perception.
• Analyzing social media data offers insights into real-world events and public sentiment.

Purpose of the Study:

• To analyze Twitter data related to COVID-19 to understand public sentiment.
• To develop and evaluate a novel sentiment classification algorithm for social media data.
• To compare the performance of the proposed algorithm against existing machine learning models.

Main Methods:

• Collected Twitter data using keywords such as COVID-19, coronavirus, deaths, new case, and recovered.
• Developed a Hybrid Heterogeneous Support Vector Machine (H-SVM) algorithm for sentiment classification.
• Evaluated the H-SVM algorithm's performance using precision, recall, F1 score, and accuracy, comparing it with Recurrent Neural Network (RNN) and Support Vector Machine (SVM).

Main Results:

• The study successfully classified Twitter sentiment regarding COVID-19 into positive, negative, and neutral categories.
• The proposed H-SVM algorithm demonstrated competitive performance metrics compared to RNN and SVM.
• Sentiment analysis of Twitter data revealed significant public anxiety and fear related to the pandemic.

Conclusions:

• Social media sentiment analysis is a valuable tool for understanding public response to health crises like COVID-19.
• The developed H-SVM algorithm shows promise for accurate sentiment classification in large-scale social media datasets.
• Understanding public sentiment through social media can inform public health strategies and interventions.