An assembly retrieval method based on the hypergraph and relationship mining
Wei Bo Li, Xidong Luo, Jie Zhang · 2024
In recent years, assembly retrieval has been a focus of research due to its potential for knowledge reuse. Many studies use pairwise distance to measure differences between assemblies, comparing the query model to each dataset model individually. However, this approach overlooks the impact of relationships among dataset models on retrieval results. This paper proposes an assembly retrieval method based on shape hypergraphs, which reflects the beneficial impact of the relationship between models on retrieval results. Firstly, a shape relationship expression model is constructed using hypergraphs to explore interrelationships among assemblies. Secondly, assembly retrieval is achieved through transductive learning on the hypergraph, avoiding direct similarity calculations. Lastly, experiments are conducted on the assembly dataset and demonstrate the effectiveness of this approach.