Exploring a Hybrid Deep Learning Framework to Automatically Discover Topic and Sentiment in COVID-19 Tweets

TL;DR

The paper addresses the need for scalable analysis of public opinion on COVID-19 by introducing a two-part framework: unsupervised topic discovery via LDA with automatic topic labeling derived from sentiment and aspect-term unigram clusters, and supervised multiclass sentiment classification using a hybrid GRU-BiLSTM model with Word2Vec embeddings and Global Average Pooling. The approach yields 14 coherent topics with labels that closely align with tweet content and achieves an average sentiment classification accuracy of about $86\%$ on the COVID-19 tweet dataset, outperforming several baseline models. Topic-based sentiment analysis is then used to map sentiment distributions to each topic, providing policymakers with actionable insights. The work demonstrates that combining topic modeling with a hybrid deep learning sentiment model can efficiently extract structured public opinion from large-scale social media data, with potential extensions to other domains and languages.

Abstract

COVID-19 has created a major public health problem worldwide and other problems such as economic crisis, unemployment, mental distress, etc. The pandemic is deadly in the world and involves many people not only with infection but also with problems, stress, wonder, fear, resentment, and hatred. Twitter is a highly influential social media platform and a significant source of health-related information, news, opinion and public sentiment where information is shared by both citizens and government sources. Therefore an effective analysis of COVID-19 tweets is essential for policymakers to make wise decisions. However, it is challenging to identify interesting and useful content from major streams of text to understand people's feelings about the important topics of the COVID-19 tweets. In this paper, we propose a new \textit{framework} for analyzing topic-based sentiments by extracting key topics with significant labels and classifying positive, negative, or neutral tweets on each topic to quickly find common topics of public opinion and COVID-19-related attitudes. While building our model, we take into account hybridization of BiLSTM and GRU structures for sentiment analysis to achieve our goal. The experimental results show that our topic identification method extracts better topic labels and the sentiment analysis approach using our proposed hybrid deep learning model achieves the highest accuracy compared to traditional models.

Figures (11)

• Figure 1: Architecture of the proposed framework
• Figure 2: Model summary for hybrid deep learning model.
• Figure 3: Coherence score for the number of topics.
• Figure 4: Top 20 unigrams from sentiment terms cluster of topic no. 3
• Figure 5: Top 20 unigrams from aspect terms cluster of topic no. 3