Data mining of weibo for public sentiment evolution regarding COVID-19

ABSTRACT

This article proposes a Weibo public opinion evolution model based on data mining for COVID-19. Python crawlers are used to collect Weibo public content, and the evolution process is classified into four phases according to the popularity of public opinion. Naive Bayes classifier is employed for sentiment analysis, and data visualization method is adopted to explore the frequent word. Regional heat characteristics at each phase, the temporal and spatial laws of public opinion are discussed accordingly. The analysis results show that through the data mining of Weibo public opinion, the evolution pattern and hot content of each phase can be identified. This study suggests the government should focus on the evolution of public opinion and take effective guidance measures timely. © 2020 IEEE.