Abstract: With the increasing penetration of the Internet in the public domain,local or single emergencies often extend to the public space due to the intervention of social media,and become a public issue widely concerned and participated by the public.The emotional accumulation and contagion of social media users may influence the trend of online public opinion.It is very important to judge the types of emotions in social media public opinion,analyze the characteristics of emotional transmission and then verify the factors that affect emotional transmission.Taking the "Kunshan Case" as the starting point,using the methods of emotion mining and social network analysis,this paper makes a coarse-grained and fine-grained identification of the network emotion in the public opinion of unexpected events and the depth analysis of network phantom and cascade.It is found that in the Weibo field,the emotional tendency of the public opinion of "Kunshan Case" is obvious,and the emotional distribution of "good" emotion is the most significant;the main path of emotional transmission is broadcast transmission,and emotional contagion shows obvious characteristics of convergence.The development trend of events and emotional polarity are both related factors of emotional contagion,and with the increase of the depth of cascade and the deepening of discussion,emotion gradually tends to be rational,which shows the "self-purification" function of Weibo as a public opinion field.

Key words: emotional contagion, complex network, broadcast network structure, structural viral index