Online Social Media Public Emotions Mining during a Public Health Emergency
Online social media is a critical channel for regulators to disseminate information and for the public to express feelings during public health emergencies.Understanding public's emotions in response to different news can help the public health department develop effective risk communication strategies.This study applies natural language processing methods to categorize social media news,and identify public emotions expressed in the corresponding news comments by proposing a method for mining public emotions on online social media during public health emergencies.Specifically,taking the recent COVID-19 pandemic as a case,this study adopts word embed-ding and clustering to analyze the epidemic-related news from three representative official Weibo accounts,and identifies eight categories of the news contents,including official news release,domestic epidemic updates,emotional support,transportation notification for track-ing,treatment information,etc.Based on the Plutchick's emotion framework,this study builds a discriminant model to classify and rate emotions expressed by a Weibo comment through crowdsourcing questionnaires and an emotion dictionary.Furthermore,this study analy-zes the impact of different news contents on public emotions and the correlation between emotions during different stages of the epidemic life cycle(the prodromal,outbreak,containment,and recovery stages).The study reveals the effectiveness of emphasizing certain risk information to nudge public emotions and increase risk awareness,providing empirical basis for risk communication strategy analysis.
public health emergencyonline social medianews content categorizationemotion identificationrisk communication
NETL
NSTL
万方数据
