COVID-19 Public Sentiment Insights and Machine Learning for Tweets Classification

Abstract

Along with the Coronavirus pandemic, another crisis has manifested itself in the form of mass fear and panic phenomena, fueled by incomplete and often inaccurate information. There is therefore a tremendous need to address and better understand COVID-19’s informational crisis and gauge public sentiment, so that appropriate messaging and policy decisions can be implemented. In this research article, we identify public sentiment associated with the pandemic using Coronavirus specific Tweets and R statistical software, along with its sentiment analysis packages. We demonstrate insights into the progress of fear-sentiment over time as COVID-19 approached peak levels in the United States, using descriptive textual analytics supported by necessary textual data visualizations. Furthermore, we provide a methodological overview of two essential machine learning (ML) classification methods, in the context of textual analytics, and compare their effectiveness in classifying Coronavirus Tweets of varying lengths. We observe a strong classification accuracy of 91% for short Tweets, with the Naïve Bayes method. We also observe that the logistic regression classification method provides a reasonable accuracy of 74% with shorter Tweets, and both methods showed relatively weaker performance for longer Tweets. This research provides insights into Coronavirus fear sentiment progression, and outlines associated methods, implications, limitations and opportunities.

Feeling Positive About Reopening? New Normal Scenarios From COVID-19 US Reopen Sentiment Analytics

“…Past research has illustrated the dramatic growth in public fear sentiment using textual analytics for identifying dominant sentiments in SCoV2 tweets [24]. Public fear sentiment was driven by a number of alarming facts as described in the subsections below.…”

“…In the US, concurrent to the physical healthcare problem, extant research observes the mass fear sentiment about Coronavirus and COVID-19 phenomena to be on a significant growth curve from the time the study started tracking the sentiment, in the February and climbing steeply towards March, 2020 [24]. According to the Center for Disease Control and Prevention (CDC), this was also around the same time that a massive number of people started seeking medical attention for COVID-19 conditions [35].…”

“…We grouped tweets into two technology-user classes: iPhone users and Android users, to explore potential behavioral aspects. Extant research supports such a classification of technology users for analyzing sentiment, psychological factors, individual differences and technology switching [24], [47]- [49]. We identified 1794 Twitter for iPhone users, and 621 Twitter for Android users, in our dataset and ignored smaller classes, such as users of web client technologies.…”

“…Extant research has demonstrated the validity of using tweets sentiment analysis to understand human behavior [24]. Sentiment analysis of the reopening tweets data has indicated dominant sentiment trends for trust and anticipation, and a larger proportion of positive sentiment.…”

“…Fear became a prominent public sentiment as awareness of the seriousness and devastating effects of SCoV2 pandemic increased [24]. This sentiment was not unjustified, due to diverse risks associated with the pandemic.…”