The node importance evaluation method based on graph convolution in multilayer heterogeneous networks

Node importance evaluation is a hot issue in complex network analysis. Existing node methods are mainly oriented to homogeneous networks, which ignore the heterogeneity of types and edges. We propose an MLN critical method based on graph convolution. In this paper, we generate feature matrix nodes. Considering diversity network, design adapted sampling meta path. An embedding model constructed convolutional (MGC). Besides, negative technique used complete MGC training. Metrics by combining vectors local structural features evaluate node's importance. The experimental results show that proposed has better accuracy than K-Shell algorithm (K-Shell), K-shell-based gravity ranking (KSDG), Page Rank (PR), influence maximization (IMNE) information entropy (ERM).