An Empirical Study Of Emotional Contagion Based On The Vent Dataset

With the continuous development of Internet technology,people's daily communication methods have also changed to a certain extent.More people are beginning to try to use online chat software as social media for a new round of making friends.However,although Internet technology has broadened the communication methods of the current group,it is also due to a certain degree of rapid information dissemination,accompanied by a certain degree of various emotional dissemination.The fast-paced way of making friends with Internet technology has also accelerated the promotion of its accompanying emotions,which is particularly likely to cause group emotions among netizens.Among the many group emotions,negative emotions have the most significant influence.If they cannot be discovered and controlled in time,they will certainly cause harm to the society to a certain extent,and it is more likely to cause certain group problems.It even derives into secondary social problems.The impact of emotional infection on the Internet is more serious,and the existence of it is more popular.This article selects the Vent data set as the research object of this emotional infection.Because users on Vent,a semi-anonymous social platform,post personal emotions anonymously,the collected data is more authentic and highly targeted.The data set is a public data set and the data is complete,which meets the needs of personalized analysis.Then the basic method of processing the Vent data set is introduced;then,the data set is preprocessed,the commonly used emotion categories are checked,and the user network structure diagram and user activity diagram are established.Through the user network graph,it can be found that a small number of users have a large number of friends,most users have a small number of friends,and the network of friends is sparse.From the perspective of relevance analysis,the Vent data set was used to count the crowd's mood fluctuations in the Vent data set,and the relevant news during the time period was analyzed using statistical measurement methods.Studies have shown that new media reports can cause emotional fluctuations among online netizens to a certain extent.At the same time,due to the special nature of the community,people who are prone to emotional fluctuations tend to drive people in the same community,causing a certain degree of emotional infection.By comparing and using the deepwalk algorithm,the complex social relationship network and propagation time in the original data set are vectorized expression,and the processed data is brought into the support vector machine,logistic regression and random forest models for calculation,and the best is selected.For the optimal parameter group,the training set and the test set are carried out according to the ratio of 7:3.Then select indicators to evaluate the quality of the model.After adjusting the parameters,select the optimal parameter group.Then the data is brought into the model,and the validity is tested,and the optimal model is obtained through comparative research.Through research,it is found that the comparison effect of various indicators of support vector machine is significant.Emotional infection classification based on support vector machine uses empirical analysis of total annual data summary.Finally,as an emotional infection classifier,the support vector machine can effectively classify all kinds of people in social networks,and divide them into ordinary emotional susceptible and negative emotional susceptible.Then you can use the classifier to check the extent of the negative emotion spread.It provides a new reference direction for the government and related departments,which has certain theoretical and practical significance.