Research On Visual Analysis Of Emotion Of Education Public Opinion Based On Web Data

Education is the foundation of national economy and people’s livelihood.With the accelerating development of network,Internet has been a new platform accumulating public opinion and expressing opinions for the civils.Therefore,research of emotion of network public opinion of education has significant values on the reflection of public attitudes and the in-depth understanding of the status quo of education.In the age of big data,the data visualization technology is a powerful tool to analyze data,and it can translate invisible and complicated data to visible forms like colors,shapes,textures and symbols using graphics and image technologies.However,there are a lot of users with analysis requirements of public opinion of education in the current society.And some of them know nothing about the expertise of education and data analysis.Therefore,how to demonstrate emotion characteristics of network public opinion of education by designing visual graphs,and how to help users carry out effectively an analysis are challenges for the current research on network public opinion.Based on the analysis requirements of target users,the new visualization scheme based on plant metaphors is proposed and two visualization systems for different users’ requirements are developed.The systems can visually display emotional characteristics,emotional propagations and emotional development trends of network public opinion of education.First,this paper proposes a method of emotion recognition based on the emotion dictionary.It combines a vocabulary of adverbs and a vocabulary of negative words,and uses the word2vec model to optimize the emotion dictionary,which improves the accuracy of emotion recognition.Besides,an emotion communication model of network public opinion of education is proposed based on the information ecological chains to define the types of emotion propagations and the efficiency of emotional propagations.Second,this paper designs and implements an visualization system based on the users’ requirements of emotion clustering.To start with,we conclude the requirements of emotion analysis of education by discussions with different types of users.Then,based on the garden metaphor,the emotion clustering map is designed to show the emotion tendencies from different data granularity.The system combines the emotion propagation view and the matrix view to support users to conduct comparative analysis of emotional propagation characteristics among different media platforms.The system can help users effectively analyze the emotion characteristics of public opinion of education and select the suitable media platforms through case studies of web data of educational public opinion.Moreover,this paper develops an visualization analysis system based on the users’requirements of evolutionary emotion.In the system,the emotion time series view is designed with the vine metaphor,focusing on the evolution of emotion over time from different data granularity.And the grid-based comparison view is designed to assist the user in analyzing the differences of emotion between the articles and the commentaries.Besides,the system combines a contrastive view of the multimedia platforms’emotional propagation and a timeline view to show the development stages of emotion of a single event.It can show users the emotion tendencies from different data granularity.Finally,the case studies of web data of public opinion of education substantiate the effectiveness and availability of the systems.