Research On Multi Label Classification Of Social Network Based On Large Scale Data

With the rapid development of the Internet,social networking is also deep into everyone's life.In recent years,the social network has attracted more and more attention,and there are many networks such as QQ and Facebook which have a large number of users.The study of the properties of these social networks attracts more and more researchers to focus on improving the existing systems and designing new social network applications.Multi label classification is an important application in the social network.With the increasing popularity of label classification,multi label classification of network data has become a research hotspot in the field of data mining.The main work of this paper is as follows:(1)Multi label analysis based on real data.Each piece of information of users in a social network can be regard as a label.Then,through the analysis of several social networks,we take the QQ social network with a large number of users,the most widely used data as the research object,and its social background attribute is systematic analyzed and the distribution of label information users of the social network is summarized,based on the existing data to assess the structure of the QQ real network.This provides us a detailed understanding on social networks.(2)As for multi label classification algorithm applied to large-scale social networks,and network data problems brought by high time complexity of the proposed multi label classification algorithm for large-scale social network,based on the structural properties of social networks,we proposed a multi target core network to reduce the cost of time.The experimental results are compared with the existing multi label classification methods and show that our proposed method can help to predict more accurate and effective labels in the network environment.(3)Combined with the classical algorithm and our algorithm,we designed a tool for multi label classification,and the purpose of this tool is to facilitate the calculation of multi label classification algorithm and intuitive show the results of algorithms.