Research On User Influence Of Weibo Specific Topic Domain Based On Interaction Relationship

User influence analysis is one of the basic issues that need to be studied in social networks,and has been widely studied by researchers domestic and foreign.Most of the existing user influence research only considers one behavior of users in social networks,or separates different behaviors and measures user influence through simple weighting,without considering the correlation between different types of user behavior and the correlation between behavior and topic content.In addition,most of the traditional weibo hot topic analysis focuses on the whole topic event,but ignores the research on the subtopic discovery in hot topics.In view of the above problems,the specific work content and innovation points of this thesis are as follows:(1)In view of the current lack of research on the subtopic discovery of micro-blog hot topics,this thesis proposes the micro-blog subtopic discovery algorithm based on time sequence,and adopts the improved subtopic correlation degree algorithm to measure the changes of subtopics in adjacent time slices,so as to accurately measure the similarity between subtopics.Firstly,text preprocessing is performed on the dataset,including Chinese word segmentation and part-of-speech tagging.Then the time series are divided into time slices,and text is divided according to different time slices.Then,the LDA topic model is used to extract subtopics for the text set of each time slice.Finally,the extracted subtopics are calculated by topic correlation degree to accurately obtain subtopics in different time slices of hot topics.The implementation shows that the improved sub-topic association degree algorithm can accurately measure the similarity between topics in adjacent time slices,and can more accurately analyze the topic evolution of micro-blog.(2)In view of the problem that existing user influence analysis methods generally only consider the single interaction relationship between users,this thesis proposes an improved user influence algorithm that integrates user interaction relationship.Firstly,according to the interactive relationship between users in a certain topic of microblogand the concern relationship between users,the model of micro-blog information interaction network is constructed.Then,the user comment strength,forwarding strength and mention strength are calculated through the microblog interaction information model.The interaction strength factor is calculated according to the strength of these three relationships.Finally,the interaction intensity factor is introduced to construct the user influence model based on the interaction relationship at the topic level of weibo,so as to analyze the user influence in the hot topic.The experimental results based on Twitter real data show that compared with other similar algorithms,the improved algorithm algorithm has obvious advantages in Accuracy,Recall rate and F value.In summary,based on the topicality of Weibo users' influence,and the different behaviors reflecting the influence of Weibo users from different angles,this thesis studies the methods of subtopic discovery method and influence analysis method based on interaction relationship.The research has important application value in the fields of advertisement placement,public opinion monitoring,viral marketing and emergency warning.