SAFAR-Onto: A Halal Tourism Ontology Developed Through Hybrid AI–Human Engineering

Halal tourism is a fast-growing market. However, information about halal services, certifications, and facilities is scattered across many sources. Existing tourism ontologies model general tourism concepts such as destinations, activities, and accommodations, but they do not represent Islamic requirements such as prayer facilities, halal certification, or dietary compliance. Manual ontology development is slow. Large language models (LLMs) can speed up concept extraction and hierarchy proposal, but they cannot decide religious or cultural matters. In this article, we present SAFAR-Onto, an OWL 2 DL ontology for halal tourism. We develop it through a hybrid AI–human process across four phases: knowledge acquisition, conceptualization, formalization, and dual validation. The process adapts METHONTOLOGY and Simple Knowledge Engineering. LLMs assist with concept extraction and dimensional analysis. An Islamic studies validation gate routes culturally sensitive decisions to human experts. We evaluate the ontology with two methods: expert review across five quality dimensions, and the OQuaRE technical metrics. SAFAR-Onto contains 10 main classes, 75 subclasses, 201 object properties, and 200 data properties, built from 67 validated domain concepts. Six experts from tourism management, computer science, and Islamic studies rated the ontology between 4.0 and 4.5 on a five-point Likert scale, with accuracy highest at 4.5. Fourteen of 18 OQuaRE metrics scored Excellent or Very Good. We also report a verified mapping from expert feedback to ontology revisions. The ontology is publicly available at https://w3id.org/def/SAFAROnto under a CC-BY-4.0 license. We hypothesize that the development process can transfer to adjacent culturally specific domains such as Islamic finance and kosher food systems.