Identifying different community members in complex networks based on topology potential
Frontiers of Computer Science in China
| Hi-index | 0.01 |
Evaluating nodes importance in the network is an important research topic in the fields of complex networks, social network analysis and graph-based data mining. Enlightened by the knowledge of physical field, this paper proposes a topological ranking algorithm based on data field theory. Its basic idea is that each node in the network can be viewed as a material particle which creates a potential field around itself and the interaction of all nodes forms a topological field over the network. By defining and computing the topological potential score of each node, we can obtain a more accurate global ranking which can reflect nodes importance in the network. The theory and experimental results on toy and real-world networks indicate that when each node only influences its neighbors, the topological ranking is consistent with the degree ranking; as the node's influence spreads, the topological ranking turns out to have similar linear relevance with PageRank sort; when the influence reaches the diameter of the network, the topological ranking is close to the closeness ranking this time. Hence, this method not only provides a general framework for some classic ranking measures, but is also able to obtain an accurate ranking of vital nodes by optimizing correlative parameters, which can reveal the position differences of network topology. Keywords-- node importance; data field; graph; complex networks