Construction and Application of Knowledge Graph for Electronic Assembly Process Standards

Addressing the challenges of knowledge acquisition, sharing, and updating in the management of knowledge for Electronic Assembly Process standards, this paper proposes an intelligent management method based on a knowledge graph. This knowledge graph provides functions such as knowledge retrieval, association, recommendation, and management, significantly promoting the sharing and collaboration of knowledge in the Electronic Assembly Process standards and enhancing the efficiency of knowledge management in this field. In terms of knowledge extraction, existing entity-relation extraction methods often perform unnecessary operations to extract subject-object pairs for relation types not included in the text, leading to a decrease in efficiency. To address this problem, this paper introduces a joint entity relation extraction model based on relation prediction and Global Pointers. By preemptively determining the relations present in the text and employing Global Pointers for subject and object recognition and alignment, the performance of triple extraction is improved. Experimental results on the Electronic Assembly Process requirements dataset demonstrate that our model has improved the $F_{1}$ score by $0.49 \%$ and $2.92 \%$ compared to Casrel and PRGC.