Abstract:

The paper targets the rough information feature of the social network in Web environment, which is that the Web data has a tremendous amount of junk content and garbage link and lots of information is incomplete, missed and repeatable. Firstly, based on the proposed the attribute graph model and combing the advantages of rough set theory solving incomplete information, the rough vertex attribute graph and the rough edges graph are proposed. Furthermore, the rough attribute graph that can describe the complex Web of incomplete social network information and the dynamic changes of links is given so that the Web social network structure is analyzed in an even better fashion. Secondly, the rough characteristics of the rough attributes graph is analyzed in order to give the concepts of rough vertex accuracy, rough edge accuracy, and rough graph accuracy and obtain the conclusion that the accuracy of rough attributes graph is related to the division level of vertex and edge set property. In other word, people's cognition of graph is closely related to the accuracy of graph. Finally, the paper authors are queried to get the rough graph in Chinese HowNet. By constantly adding the vertex attributes and dividing the graph vertex, the author's cooperation relations figure is excavated, thus demonstrating that the rough attribute graph in social network analysis is in line with the cognitive processes of people.

Key words: web social network;attributes graph;rough attribute graph;graph precision