ALE semantic interoperability experiments (E1–E7): code, ontology, synthetic test data and aggregated results

Abstract

Replication package for the book chapter Semantic Interoperability of Heterogeneous RDF-Based Knowledge Models: Expressiveness Assessment, Loss-Aware Mapping and Integration in an Agricultural Advisory Environment by J. Rogushina, A. Gladun, S. Pryima, O. Anishchenko and A. Savochka (under review for Advanced Computing Applications, Springer). The package contains: the programs of experiments E1–E7: mapping of OWL 2 ontologies to Semantic MediaWiki (direct mapping M0 and loss-aware mapping M1) and back, competency questions, n-ary registered uses of plant protection products, personalised projections, construction of LLM context from the knowledge graph (S2) with answer control (G2), evaluation with real LLMs (llama3.1:8b, gemma3:4b) against BM25, dense and hybrid retrieval baselines, and Semantic MediaWiki scalability; the working draft of the ALE ontology (1.0.0-draft) used in the experiments, with synthetic test individuals and SPARQL competency questions; open data of the State Register of pesticides and agrochemicals of Ukraine (2008–2024), redistributed unchanged; aggregated results and figures. Records extracted from the copyrighted handbook «Пестициди» (2019) are not included. The package contains the extraction programs and SHA-256 checksums, so these data can be regenerated and verified from one's own copy of the handbook. Synthetic data are marked as such. See README.md. Licences: code GPL-2.0-or-later; ontology and data derived from it CC BY-SA 4.0; other results and documentation CC BY 4.0; register data — open data of the Ministry of Economy of Ukraine (see LICENSES.md). The published ALE ontology, version 1.0.0, is a separate record: https://doi.org/10.5281/zenodo.23643625. Funding: Ministry of Education and Science of Ukraine, research project "Development of an Intelligent Support Environment for Agricultural Advisory Services Based on the Integration of Semantic Wiki Technologies and Artificial Intelligence" (state registration No. 0126U000765).