Identifying Influential Nodes In Online Social Networks

As the development of computer technology and information technology, online social network has attracted the huge concern of people. Online social network is refers to a type of network which formed by people who socialize, and contact and collaborate with others via computer and network as intermediary, such as blog, and micro-Blog, and instant communications.Identifying influential nodes in online social networks has a very important role for theoretical significance and application value, such as accelerating information propagation, controlling rumors and diseases, designing search engines, and understanding hierarchical organization of social and biological networks. Known methods range from node centralities, such as degree, closeness and betweeness, to diffusion-based processes, like PageRank and LeaderRank. Duan-Bing Chen[1] and other researchers have pointed that the local community structure plays a negative effect to the information diffusion. Although it exists the reason about local community structure which will have a limitation for the information to spread more widely, the structure holes of social network could make up this deficiency. Therefore, we think the most influential nodes in the network structure should not only consist of the important nodes which located inside the community, but also the structure holes among the communities.For the above reasons, an algorithm about identifying the structure holes in the information diffusion network has been proposed. The algorithm considers not only the community structure in the information diffusion network, but also the significant nodes inside the community structure which makes the process of identifying structure holes more efficiently. And according to the existence of structure holes and community inner important nodes, we also bring up the definition of influential nodes in the information diffusion network. Later, we experiment the algorithm on the dataset of Sina microblog and compare our algorithm with PageRank and PathCount and the results shows that our algorithm performs effectively.