BSIN: A behavior schema of information networks based on approximate bisimulation

Information networks are increasingly important in practice, of which the escalating complexity is gradually impeding efficiency of data mining. To address this issue, a novel network schema called Behavior Schema of Information Networks (BSIN) is proposed. In BSIN, this paper defines the behavior of nodes as connected paths; proposes a novel metric to distinguish behavior difference; forms metric spaces for different node types based on the novel metric; and introduces approximate bisimulation into obtaining quotient sets for metric spaces. The biggest highlight of BSIN is the ability to directly obtain a high-efficiency network based on approximate bisimulation, rather than reduce the existed information network. It provides an effective representation for information networks, and the resulting novel network has a relatively simple structure that efficiently expresses semantic information compared to current network representations. Theoretical analysis of connected paths between the original network and the obtained one demonstrates that errors are controllable, and semantic information is approximately retained. Case studies show that BSIN yields a relatively simple network and is highly cost-effective.