Research on Topic-emotion Graph Construction and Application of Online Public Opinion of Emergencies
[Purpose/significance]Grasping accurately the topic evolution of public opinion and effectively perceive netizen emotions,monitoring the development trend of public opinion and enhancing the early warning ability for public opinion crisis can provide new ideas and methods for managing public opinion.[Method/process]Aiming at the high difficulty in perceiving public opinion topics and the overall emotions of netizen in online opinion emergencies,as well as the lack of traceability and interaction analysis,on the basis of using sentiment lexicon to calculate netizen emotions and BERTopic model to mine the topics of netizen comments,this paper constructs a topic-emotion graph including three types of entities:events,comments and users,and their relationships,and applies Neo4j graph database to store nodes and relationships,and uses Cypher language for querying and visual presentation.[Result/conclusion]Based on the crawled dataset of online public opinion comments of emergencies,the paper achieves the design and construction of the topic-emotion graph,which can be effectively applied in overall emotional perception,user emotion tracing,and topic emotion perception.It enables interactive analysis of public opinion topics and netizen emotions,deeply explores the factors that affect netizen emotions in public opinion,and provides valuable reference for public opinion governance.
online public opiniontopic-emotion graphpublic opinion evolutionNeo4j
万方数据
维普
