Knowledge graph-based enterprise employment risk early warning method

Abstract

The present invention discloses a knowledge graph-based enterprise employment risk early warning method, including: acquiring employee-position matching data, system execution records, and team collaboration logs from an enterprise’s internal database and external compliance information sources through a data collection module to obtain a multi-subject associated dataset; performing node representation learning on the multi-subject associated dataset using a graph neural network module to determine an associated graph structure; extracting subgraphs from the associated graph structure using a subgraph extraction module to obtain a risk source subset; judging a dynamic transmission trend from the risk source subset using a trend judgment module, and tracking a transmission path using the dynamic transmission trend to obtain a risk amplification chain; calculating the path length and weight according to the risk amplification chain to determine a potential hazard scope; comparing the potential hazard scope with a current dataset through a data comparison module to obtain an early warning mechanism activation signal.

Images