Design And Implementation Of Group Emotion Analysis And Visualization System Of Micro-blog

With the advent of the mobile Internet era,the real-time topic of online user has exploded in the network.How to obtain valuable "information drops" and conduct in-depth analysis in the sea of large data of network text is a hot topic of current research.Especially,the hot issues of weibo online reviews of sentiment analysis,can help us to timely and effectively understand the group emotion and the trend of public opinion,which has important research value and broad application prospects.Based on sina weibo,this paper conducts in-depth research on text emotion classification and group emotion analysis,and designs a group emotion visualization analysis system for news hot events.First of all,the characteristics of microblog text are deeply analyzed,and a text emotion classification model based on feature fusion and a text emotion classification model based on multi-classifier are proposed.LSTM,GRU,CNN and other neural network models are used to carry out feature training based on 45,000 microblogs.Meanwhile,attention mechanism is introduced to carry out feature weighting.The experimental results show that the proposed method has a higher accuracy rate of 86% and 85% respectively,compared with the maximum accuracy of 75% for single neural network.Secondly,the group emotion analysis was carried out for the real-time hot events of sina weibo platform.Through the research,we selecte the optimal performance of the text emotion classification model based on feature fusion and "thad in Korea" and "exercises" south China sea two hot issues for emotion classification and sentiment analysis,the experimental results show that Internet users emotional expression is more and more diverse and rational,and a growing sense of belonging and identity to the nation.Finally,a group emotion visual analysis system for news hot events is designed and implemented.This system can analyze the emotional distribution characteristics of netizens in different events based on the number of topics and the change trend of activity degree of weibo users in different types of events.The experimental results show that the group emotion in hot events has obvious time periodicity and distribution stability.