Study on the construction of knowledge graph for diseases and pests of citrus

doi:10.16178/j.issn.0528-9017.20231218

Abstract

Abstract:

In traditional cultivation of citrus, farmers lack abundant professional agricultural knowledge and technology. Information technology can digitize the knowledge of pest and disease control of citrus. However, current knowledge management technologies have some common deficiencies in terms of efficiency, scalability, and applicability. Knowledge graph provides a new approach to knowledge management, which is a more flexible method. Considering the insufficient detailed characterization of entities and relationships related to pest and disease control of citrus in existing agricultural knowledge graphs, this study defines entity types and relationship categories for pest and disease knowledge of citrus. A total of 10 entity categories and 10 entity relationships were identified, and a large language model was used for extraction and deduplication, addressing 16 common deficiencies. Finally, the knowledge graph was stored and visualized using the Neo4j database. The discussion focused on how knowledge graph for pest and disease of citrus can provide underlying technical support for downstream tasks such as precise pest and disease information inquiry and intelligent-assisted diagnosis.

Key words: pest and disease of citrus, knowledge graph, Neo4j, large language model

CLC Number:

S436.66

ZHU Yihang, ZHANG Xiaomin, ZHOU Tianqi, RAO Xiuqin. Study on the construction of knowledge graph for diseases and pests of citrus[J]. Journal of Zhejiang Agricultural Sciences, 2025, 66(3): 687-691.

Figures/Tables 6

Fig.1 Framework diagram of knowledge graph construction

Fig.2 Ontology modeling in the graph domain

Table 1 Pest and disease dataset of citrus

关系类型	数量/个	实体类别	数量/个
分布	72	症状	176
预防	85	特征	283
治疗	99	地理位置	46
危害	46	果树位置	24
关注	18	规律特点	205
表现	223	时间	41
发生	233	预防方法	60
特征显示	308	治疗方法	84
暴发	48	病虫种类	3
防治	29	危害等级	9

Fig.3 Visualization of knowledge graph of citrus disease and pest

Fig.4 Information inquiry of pest and disease

Fig.5 Inquiry of the hazards location of pest and disease

References 13

[1]	尹怀中, 傅中渝, 彭昌家, 等. 柑橘病虫害全程绿色防控技术集成试验示范效果[J]. 中国植保导刊, 2018, 38(3): 51-57.
[2]	刘杰帆, 林铭佳, 张建桃, 等. 基于CiteSpace可视化分析的柑橘黄龙病研究进展[J]. 果树学报, 2023, 40(4): 771-787.
[3]	邓明学. 以控制木虱为重点的柑橘黄龙病综合防治技术研究[J]. 植物保护, 2006, 32(6): 147-149.
[4]	WANG Y, WANG Y, WANG J, et al. An ontology-based approach to integration of hilly citrus production knowledge[J]. Computers and Electronics in Agriculture, 2015, 113: 24-43.
[5]	GUO M, ZHANG J Y, GENG N, et al. Chinese relation extraction of apple diseases and pests based on BERT and entity information[M]// Knowledge Science, Engineering and Management. Cham: Springer International Publishing, 2022: 579-592.
[6]	唐闻涛, 胡泽林. 农业知识图谱研究综述[J]. 计算机工程与应用, 2024, 60(2): 63-76.
[7]	吴赛赛, 周爱莲, 谢能付, 等. 基于深度学习的作物病虫害可视化知识图谱构建[J]. 农业工程学报, 2020, 36(24): 177-185.
[8]	张永亮. 基于知识图谱的苹果病虫害智能问答系统研究[D]. 杨凌: 西北农林科技大学, 2022.
[9]	闫丽华. 基于知识图谱的葡萄病虫害自动问答系统[D]. 杨凌: 西北农林科技大学, 2021.
[10]	陈亚东, 鲜国建, 寇远涛, 等. 我国苹果产业知识图谱构建研究[J]. 中国农业资源与区划, 2017, 38(11): 40-45.
[11]	王栋, 周菲, 李颖芳, 等. 我国甜樱桃产业知识图谱构建研究[J]. 中国果树, 2023(1): 104-108.
[12]	杨林楠, 郜鲁涛, 林尔升, 等. 基于Android系统手机的甜玉米病虫害智能诊断系统[J]. 农业工程学报, 2012, 28(18): 163-168.
[13]	VASSILIOU G, PAPADAKIS N, KONDYLAKIS H. SummaryGPT: leveraging ChatGPT for summarizing knowledge graphs[M]//The Semantic Web:ESWC 2023 Satellite Events. Cham: Springer Nature Switzerland, 2023: 164-168.