Public Emotion Monitoring and Surveillance in Emergencies

LIU Zhi-ming,LIU Lu
IF: 2.034
2010-01-01
Systems Engineering
Abstract:The public emotions in emergencies may lead to other public events,so it is necessary to monitor and track the evolution of the public emotions in time.This paper studies the evolution of the public emotions during the emergencies.Firstly,we construct a hierarchical emotion model for emergencies and an emotion monitoring surveillance model.Secondly,we use three machine learning algorithms and three kinds of feature selection methods to study the emotion classification for Chinese microblog about the emergencies.The experimental results show the performance of SVM and IG feature selection is best,and the accuracy of the classification is fit for need,then the paper discussed two real cases and analyses the evolution of emotions.Finally we define some surveillance indexes and surveillance patterns for emergency early warning.
What problem does this paper attempt to address?