Research On Topic Drift In Social Networks

With the development of Internet technology and the popularity of mobile intelligent terminals,social networks have developed rapidly and become an important platform for information exchange.Its fission information dissemination mode promotes the rapid dissemination of information in a wide range,but at the same time it also brings a lot of topic drift.The emergence of topic drift distracts users'attention from the target topic and reduces the effectiveness of information dissemination.In order to reduce the negative impact of topic drift and effectively guide public opinion and social network marketing,it is necessary to deeply analyze and explore the phenomenon of topic drift.This paper takes topic drift in social networks as the main research object,and studies topic drift from four aspects:topic drift phenomenon identification,topic drift path analysis,key node discovery of topic drift,and cause analysis of topic drift.Taking the popular topic of Sina Weibo as an example,"Dying to Survive",this paper captures 219 658 data of Sina Weibo about the movie"Dying to Survive" from July 8,2018 to July 23,2018,and uses social network analysis method and keyword analysis method to analyze the data.The results show that when there are independent themes which are free from related themes,and independent themes account for a large proportion of all themes,for example,this paper takes 28.6%,which indicates that there is topic drift in the process of topic discussion.When there are unrelated topics in the related topics,and the proportion is relatively large,taking this article as an example,unrelated topics account for 20%of the related topics,that is to say,there is topic drift in the process of topic discussion.The results show that the inflection point of topic life cycle is the key node of topic drift;the node with more than 50%new high-frequency keywords is the key node to generate topic drift;the node with unrelated topic names,place names or other proper nouns is the key node of topic drift.In addition,the results also show that the outbreak of relevant hot issues,the emergence of opinion leaders,the initiation of micro-blog activities,spontaneous group activities are the main causes of topic drift.This paper studies topic drift from different aspects of topic drift,which makes up for the research gap in related fields,and provides theoretical reference for intervention and guidance of topic drift to a certain extent.