Research on Knowledge Transfer on Multilayer Networks Based on Link Prediction Algorithm

Abstract

Abstract Multi-layered networks have great advantages in portraying the multi-attributes of links and can describe complex real-life systems better. Link prediction and knowledge transfer in complex networks have been extensively studied, but link prediction and knowledge transfer on hierarchical networks are less of a concern. Based on the definition of hierarchical network, a random walk model including link prediction and knowledge transfer is proposed. The link prediction method is proposed from the structural similarity and knowledge compatibility, and then the knowledge transfer rules are proposed. This paper also proposes the evaluation indicators for link prediction and knowledge transfer. The experimental results by using real hierarchical networks show that the link prediction has obtained better results and the complexity has been reduced; the knowledge transfer efficiency has been improved. This study has important reference value for the development of multi-layer network theory.