Twitter sentiment analysis on coronavirus: Machine learning approach

Abstract

In machine learning, a fundamental challenge is the analysis of data to identify feelings using algorithms that allow us to determine the positive or negative emotions that people have regarding a topic. Social networks and microblogging are a valuable source of information, being mostly used to express personal points of view and thoughts. Based on this knowledge we propose a sentiment analysis of English tweets during the pandemic COVID-19 in 2020. The tweets were classified as positive or negative by applying the Logistic Regression algorithm, using this method we got a classification accuracy of 78.5%.