Abstract:

Integrating complex and large-scale agricultural ontologies into a unified framework is crucial for eliminating data silos across platforms, optimizing the standardization of agricultural knowledge representation, and enhancing information retrieval efficiency. Leveraging the inherent structural advantages of graph databases in ontology storage, this study proposes an innovative method for importing large-scale agricultural ontology data in both OBO and OWL formats into a graph database. The method first involves semantically parsing and splitting OBO ontologies, while simultaneously processing OWL ontologies through the elimination of redundant concepts and resolution of prefixed resources. To reduce storage overhead, an encoding scheme and a co-occurrence frequency-based attribute-relation filtering strategy are further designed. Finally, intelligent modeling and mapping are performed to store the ontologies within the graph database, resulting in the construction of an agricultural multi-ontology database comprising 167,887 entities and 249,603 relationships. Comparative analysis of entities and relationships demonstrates that the proposed method effectively preserves both internal ontological structures and extensive inter-ontological knowledge links. Case studies confirm that the multi-ontology parsing and integration mechanism enables intuitive and effective cross-ontology knowledge interaction. This approach facilitates the reuse and sharing of agricultural ontologies, significantly improving the standardization of agricultural information resources. The constructed integrated agricultural multi-ontology knowledge base provides a robust data foundation for semantic search, deep knowledge mining, and intelligent decision-making in agriculture.

Key words: agricultural ontology, graph database, ontology parsing, knowledge integration