Node capability dependency importance evaluation of heterogeneous target operational network

Changjiang Qin, Yi Liang, Jincai Huang, Zhong Liu

Abstract: It is very important to identify the important nodes in heterogeneous target operational network (HTON) for target selection. Traditional methods, such as centrality based methods, only consider either network structures or node features to evaluate the importance of nodes. However, these methods ignore the rich semantic information between nodes formed by complex physical and logical relations. To solve this problem, we propose a novel node capability dependency importance evaluation method considering meta-path and capability dependency, called NCDI. Meta-path is an effective method of semantic capture, different meta-paths express different semantic informations. Some nodes could support or control other nodes, inter-node dependency can also be used as an indicator of importance evaluation. Compared to the degree centrality and other typical evaluation methods, the results show that our method can sort the node importance more effectively in HTON.