Analysis And Research Of Think-tank User Portraits Based On Weibo Platform

With the rapid development and reform of the Internet,the Internet social media platform carrying people's desire to express is also developing and expanding.Represented by sina weibo,users on the platform of the weibo left a large amount of information data,are all users in the use process of representative data,not only reliable,unique also true,the data are collected through the processing,the user portrait technology for data analysis and mining,the user can be category and said,Then,a comprehensive and accurate virtual user portrait composed of multiple tags can be formed,which to a certain extent can enable platform organizations to form a deeper understanding of users,and lay a foundation for subsequent target user positioning and personalized research.In the existing researches on user portraits,it is not common to analyze the user groups of think tanks.Therefore,this paper mainly collects the data of think tank users from the microblog platform,obtains user characteristics,forms a user label system,and finally constructs the think tank user portraits based on the microblog platform.The main research is divided into the following points :(1)combing and summarizing the technical basis of data collection,processing,data mining and analysis based on the current theoretical research of user portrait.(2)Demand analysis was conducted on the user portrait of the think tank,detailed design and step construction of the construction process were carried out,and the construction ideas and specific technical methods of each step were elaborated from the three parts of data processing,label system and user portrait formation.(3)The applicability of the existing LDA topic model is improved,and the word co-occurrence model is integrated to propose the LDA-WC topic extraction model based on LDA topic extraction and word co-occurrence,and the topic generalization and extraction are carried out for the interest range of users of Micro-think tank.According to the user label system of Weibo think tank,factor clustering analysis is carried out to extract the characteristics of user labels.(4)According to the user portrait construction process mentioned above,carry out experiments and demonstrations from data collection,tag system construction,portrait visualization and other aspects,and make targeted analysis for different categories of users and put forward corresponding suggestions.Through the above research,this paper constructs the microblogging platform based thinktank user portrait,basically completed the overall process design,a picture of the micro primus library users microblogging platform think-tank,users can be divided into the research users,active users,the life,science and technology users to four categories,and for each type of user has carried on the concrete analysis,We have a clear understanding and classification of think tank users on weibo platform.By analyzing user data,this paper hopes to obtain user characteristics and have a deep understanding of users,which will lay a foundation for the networked development of think tanks,providing high-quality services,accurately locating users and exploring more application scenarios.