A Data Retrieval Method Based on AGCN-WGAN

Abstract

Traditional knowledge graph retrieval techniques ignore node relationship weights, making it difficult to achieve targeted retrieval. Therefore, a nonlinear model AGCN-WGAN is proposed to solve retrieval tasks in relational networks, fully utilizing the advantages of Graph Convolutional Networks (GCN) and Generative Adversarial Networks (GAN). Firstly, AGCN is used to capture the local topological features of a single node; In addition, the use of GAN enhances the ability of AGCN models to generate reasonable weight distribution maps, effectively extracting correlations between nodes, thereby improving the performance of the model in handling large-scale data retrieval tasks. In order to verify the effectiveness of the method, the dispatching operation data in a real business scenario of a city power grid is used for experiments. The experimental results show that the proposed data retrieval method has significantly improved accuracy compared to existing methods.