Construction and application of knowledge graph for seed quality standard documents.

Area of Science:

• Agricultural Science
• Information Science
• Computer Science

Background:

• Seed quality standards are crucial for crop cultivation supervision and agricultural development.
• The increasing volume of seed quality standard documents in China presents challenges in efficient querying and semantic association due to their unstructured nature.
• Existing knowledge representation methods are insufficient for the seed domain, hindering effective data utilization.

Purpose of the Study:

• To develop a structured knowledge representation framework for seed quality standards using a Knowledge Graph (KG).
• To address the limitations of unstructured data in seed quality standards for improved information retrieval and management.
• To provide a digital foundation for intelligent quality management within the plantation industry.

Main Methods:

• Construction of a domain-specific ontology with 7 core classes and 12 relationship types to define semantic structures.
• Implementation of a hybrid knowledge extraction strategy combining regular expressions for semi-structured data and a BERT-BiLSTM-CRF model for unstructured text.
• Development of a Knowledge Graph (KG) containing 2436 nodes and 3011 relationships, stored in Neo4j for multi-dimensional retrieval and visualization.

Main Results:

• The proposed BERT-BiLSTM-CRF model achieved a high F1-score of 91.61% in Named Entity Recognition (NER), surpassing other models.
• The constructed KG effectively organizes seed quality standard information, enabling efficient multi-dimensional retrieval and visualization.
• Significant improvements in the accuracy of standard information retrieval were demonstrated.

Conclusions:

• The proposed Knowledge Graph construction framework provides an effective solution for structuring and querying seed quality standards.
• The hybrid knowledge extraction strategy enhances the accuracy and efficiency of data processing in the agricultural domain.
• The developed KG serves as a digital foundation for advancing intelligent quality management in the plantation industry.