Social network analysis is a hot area of data mining. In social network analysis, person or team will be considered as a node and the relationship between them will be considered as the edge, such as friendships, family relations and trade relations. When a specific relationship exists between two people or teams, add an edge between them. Thus it forms a relation network, from which you can mine lots of meaningful information. Co-authorship is an important application of social network, indicates the relationship between the coauthors. Using of social network theory of the centrality, you can mine the structure of scientific collaboration and the status of individual researchers in the co-authorship networks.In this paper, we study the co-authorship based on DBLP using social network analysis theory. The main contributions of this paper are as following.First of all we do statistical analysis related to authors and publications. This shows the co-authorship on DBLP and the characteristics of the computer science field.Then we propose a new approximation algorithm, DegRan algorithm to detect the centric author. We propose a new method of choosing estimated nodes randomly according to degree and then computer the single-source shortest-path based on them. The results show that the improved approximation algorithm has better accuracy. |