Abstract:

Wheat Sharp Eyespot is a soil-borne fungal disease commonly found in China's wheat areas, which can occur throughout the entire reproductive period of wheat and has a great impact on the yield and quality of wheat in China. By constructing a Wheat Sharp Eyespot control domain ontology and modeling domain knowledge, we aim to integrate and share the knowledge in the field of Wheat Sharp Eyespot control to provide important support and guidance for agricultural decision-making and disease control. The ontology construction process for Wheat Sharp Eyespot control is proposed to meet the actual needs of Wheat Sharp Eyespot control. For the problems of low efficiency and limited expert knowledge in constructing ontologies by manual methods, this study will explore new methods for ontology construction. Special attention will be paid to the methodology of mining core concepts of the ontology to reduce the subjectivity and limitations in the construction process, so that the ontology will have a wider application potential.In this study, used the literature in the field of Wheat Sharp Eyespot control as a data source, KeyBERT keyword extraction algorithm was used to mine the core concepts of ontology, and BERT embedding and cosine similarity were used to find out the subphrases in the document that were most similar to the document itself. Hierarchical relationships between ontology concepts were extracted by hierarchical clustering, topic modeling was performed using BERTopic, Transformer and c-TF-IDF were used to create dense clusters.Finally, Protégé was used to visualize and express the ontology concepts and inter-concept relationships.In this study, the results of thematic and hierarchical clustering were analyzed and condensed to classify the ontology of Wheat Sharp Eyespot control into eight parent concepts, which were pathogenicity pattern, wheat growth period, etiology of the disease, disease area, disease extent, symptoms and control measures. According to the characteristics of the Wheat Sharp Eyespot control domain, 11 object attributes, 16 first-level data attributes, and 8 second-level data attributes were defined for the Wheat Sharp Eyespot control ontology by organizing and analyzing the associations among the parent concepts. Finally, Protégé was used to visualize and express the ontology concepts and inter-concept relationships. This study proposed a method for constructing a domain ontology for Wheat Sharp Eyespot control, described the basic method for constructing an ontology by building a corpus of Wheat Sharp Eyespot, gived a process framework for constructing a domain ontology, and described in detail the algorithms and construction tools used in the construction. The data source of this study was mainly scientific and technical literature, and the ontology can be extended in the future by further expanding the data source. The assessment part of the ontology mainly relied on the assessment of domain experts at present, and quantitative assessment can be added in the future.The Wheat Sharp Eyespot control domain ontology constructed in this study contained a more complete conceptual system of Wheat Sharp Eyespot, meeting the ontology evaluation criteria and ontology construction requirements, and can provide reference for the construction of domain ontology, and provide powerful support for knowledge discovery and downstream applications in the field of Wheat Sharp Eyespot prevention and control, such as intelligent Q&A, intelligent recommendation, and so on.

Key words: wheat sharp eyespot, prevention and treatment, domain ontology, ontology construction, extraction, hierarchical clustering