Construction and application of knowledge graph for seed quality standard documents

Seed quality standards are the essential basis for crop cultivation supervision. With the continuous development of China's standard system, the number of seed quality standard documents has increased dramatically. However, the rapid growth and unstructured nature of standard documents hinder efficient query and semantic association. To address the lack of structured knowledge representation in the seed domain, this study proposes a Knowledge Graph (KG) construction framework for seed quality standards. First, a domain-specific ontology is constructed, defining 7 core classes and 12 relationship types to standardize semantic structure. Second, a hybrid knowledge extraction strategy is implemented: regular expressions are used for semi-structured tabular data, while a BERT-BiLSTM-CRF model is employed for unstructured text. Experimental results demonstrate that the proposed model achieves an F1-score of 91.61% in Named Entity Recognition (NER), outperforming than other model. Finally, a KG containing 2436 nodes and 3011 relationships is stored in Neo4j, enabling multi-dimensional retrieval and visualization. The proposed framework significantly improves the accuracy of standard information retrieval and provides a digital foundation for intelligent quality management in the plantation industry.S.