A knowledge graph-based agricultural information architecture for sustainable decision support

A knowledge graph-based agricultural decision support

Abstract

Sustainable agricultural decision support increasingly depends on the ability to connect different kinds of knowledge rather than simply accumulate more data. Crop characteristics, soil properties, climate conditions, water requirements, pest and disease information, farm practices, regional constraints, and sustainability indicators all enter into agricultural decisions. In many existing digital agriculture systems, however, these resources are stored in separate structures and described with different terms and assumptions. This fragmentation reduces interoperability and reuse, and it also makes recommendations harder to explain. Prediction and monitoring tools are useful, but they often leave the relationships among crops, soils, environmental conditions, management practices, risks, and decision outcomes only weakly represented.

This study develops a knowledge graph-based agricultural information architecture for sustainable decision support. The architecture represents agricultural decision knowledge through connected entities such as crop, soil, climate condition, water resource, disease, pest, farm practice, region, sustainability indicator, and recommendation. These entities are linked through relations such as requires, is suitable for, increases risk of, reduces, is affected by, and is recommended for. Representing agricultural knowledge as a graph makes it possible to support structured queries, semantic integration, evidence tracing, and interpretable recommendations.

The contribution of the study is mainly architectural. It moves the discussion from isolated repositories and model-centered prediction tools toward relational knowledge infrastructures. Such an infrastructure can support crop-soil suitability analysis, irrigation advisory, disease-risk interpretation, sustainable practice recommendation, and region-specific agricultural planning. From a Management Information Systems perspective, the study treats knowledge graphs as a reusable semantic backbone for agricultural information systems. From a Computer Engineering perspective, it provides a graph-based structure for integrating heterogeneous agricultural data and supporting explainable, query-driven, and sustainability-oriented decision intelligence.