COVID-19 Public Sentiment Insights and Machine Learning for Tweets Classification
In "COVID-19 Public Sentiment Insights and Machine Learning for Tweets Classification" (Samuel et al., 2020), the authors analyze coronavirus-related Tweets using R and its sentiment-analysis packages to track public fear-sentiment as COVID-19 approached its U.S. peak. The study is motivated by the pandemic's informational crisis, in which incomplete and inaccurate information fueled mass fear and panic, underscoring the need to gauge public sentiment for messaging and policy decisions. Using descriptive textual analytics and visualizations, the authors compare two machine learning classifiers, Naive Bayes and logistic regression, for classifying Tweets of varying lengths. Naive Bayes reaches 91% accuracy on short Tweets, while logistic regression reaches a lower 74% on short Tweets; both methods perform more weakly on longer Tweets. The paper discusses the implications, limitations, and opportunities of this fear-sentiment analysis approach.
0
1
Contributors are:
Who are from:
Tags
CSCW (Computer-supported cooperative work)
Computing Sciences
Related
Ben Hammeer expresses how Kaggle helps COVID researchers
Nature tweets about a new article about AI summarizing papers to help COVID researchers sift through literature
Vaughn Cooper tweets about a COVID registry by ASMicrobiology
Connected Papers tweets about the launch of their new product, a visual tool to help scientists find relevant papers
Axios posts about AI solutions to the flood of information related to COVID research
Raleigh M. Linville tweets useful browser extensions that help improve literature searches
Rens van de Schoot tweets about new version of plugin that uses active learning to produce recent and relevant COVID-19 papers
DiscourseNet posts webinar that allows scientists to get their COVID papers (between 1000-3000 words) reviewed by a group of peers.
Bianca Kramer posts that Elsevier's COVID related content will remain free until October
Salesforce tweets their new search tool for relevant COVID-19 papers
Geoffrey Siwo tweets about crowdsourced COVID-19 gene library that helps reach consensus
Dr.Tedros Adhanom Ghebreyesus tweets about on a new COVID-19 data platform, Pool, that will be open to everybody to work on the solution
Jen Oyler-Yaniv tweets about coming together with researchers from other universities to learn
Announcement of Cross-Publisher Collaboration for COVID-19 Peer Review
COVID-19 Public Sentiment Insights and Machine Learning for Tweets Classification
Tim Plante tweets his solution on how he helped others template their scientific papers