Analysis Of Emotional Communication Of Weibo Users Based On Neural Networks

Emotion is a crucial psychological phenomenon since it drives an individual in an advanced way.Emotion is not only contained in people's psychological activities,but also reflected in people's behavior and expression.With the advancement of the society and the Internet,various social networks have gradually become an important platform for the public to express their opinions.The hidden emotions in comments also spread among netizens while they express their view and stance.Emotion diffusion is a common phenomenon,and accumulative negative emotions in a group can result in disturbance online or offline,even with grave consequences.At present,qualitative analysis is mostly used in the study of Internet users' emotions in communication studies.At the same time,researchers' research on the evolution stage of public opinion does not highlight users' emotions.Therefore,in order to better study the emotional characteristics of users in hot Internet events,this article combines communication with the convolutional neural network sentiment analysis model in the computer field.As a case study,using Sina Weibo as the research platform,the following studies were conducted:First,it has sorted out the domestic and foreign literature,summarized the current status of research at home and abroad,and laid a theoretical foundation for the subsequent research.Second,it has took the "Yuya be domestic violence" incident as a research case and used the python to collect the Weibo data;third.It has used convolutional neural network to analysis models in the computer field to perform sentiment analysis and visualized the Weibo data.Fourth,it has divided the user's emotional evolution stages,and divided the User emotional characteristics were analyzed.Through the above studies,the following conclusions are reached: First,it has compared with the evolutionary stage which is simply divided from the popularity of Weibo reviews,the emotional evolution stage is different in fluctuation amplitude and fluctuation time.The Weibo public opinion evolution model selected in this paper The stages of public opinion evolution can be more scientifically divided.Second,the evolution process of Internet hotspot events is staged.This process can be divided into the budding stage,the erupting stage,the repetitive stage,and the fading stage.Third,the subject words in different stages can represent the trend of Weibo users 'emotions at this stage.The influence of the processing attitude of the media,opinion leaders and related departments on Weibo users' emotions cannot be ignored.Based on the research conclusions,this article proposes coping strategies for relevant departments,including: clarifying the types of incidents,building a network monitoring system,focusing on the role of opinion leaders,and conducting targeted user emotion guidance strategies for different development stages.