A Method for Evaluating Key Nodes in Heterogeneous Social Networks Using Deep Reinforcement Learning

Abstract

The invention discloses a method for evaluating key nodes of a heterogeneous social network by using deep reinforcement learning. The method first constructs a multi-layer directed network model according to the heterogeneity of the heterogeneous social network, and according to whether there is an interactive relationship in the real network Determine the connection edges between user nodes, determine the connection edges between the attribute nodes of the same layer according to the attribute similarity, and define the meta-path according to the in-degree, out-degree and transitivity of the nodes. Secondly, construct a new network embedding framework based on the deep reinforcement learning mechanism compound relation element path, and encode the multi-layer directed network into a low-dimensional embedding vector; then, construct the LBSN (Location-based Social Network, based on the improved graph neural network) location social network) key node evaluation model, and the ranking of the output node importance. The present invention can effectively improve the comprehensiveness, accuracy and efficiency of evaluating key nodes of heterogeneous social networks.

Description

technical field [0001] The invention relates to the technical field of key nodes of heterogeneous social networks, in particular to a method for evaluating key nodes of heterogeneous social networks by using deep reinforcement learning. Background technique [0002] The vigorous development of large-scale social networks at home and abroad has provided a large amount of reliable data support for the research of social networks. The research on the key nodes of social networks is one of the key issues in social network analysis, and the main theoretical framework at this stage is for single-layer network structures. The single-layer network structure is a simplified social network framework, and it is difficult to accurately describe the state of interdependence and interrelationship among different types of nodes in heterogeneous social networks. [0003] How to accurately describe the network structure of LBSN (Location-based Social Network, location-based social network),...

AI Technical Summary

Problems solved by technology

Not from the perspective of information transmission, local and global information, resulting in limitations such as low accuracy of evaluation results and weak generalization ability

Images