Ontology-augmented health recommendation
Ontology-augmented personal healthcare service recommendation is a course project for an ontology-engineering class at Chiang Mai University. It models healthcare in Thailand (diseases, symptoms, treatments, specialties, facilities and insurance) as an OWL 2 DL ontology, so a reasoner can recommend where a patient should go and what their insurance covers.
The ontology
The Thai Healthcare Service Ontology is written in OWL 2 DL and published in OWL, Turtle and JSON-LD. Its README describes:
- 85 classes covering diseases, symptoms, treatments and medical specialties;
- real Thai hospitals and clinics with location data, among 244 individuals;
- insurance: providers, coverage plans and cost calculation, including the national Gold Card scheme;
- Thai and English labels for every entity;
- 38 SWRL rules rules for detecting emergencies, recommending specialties and working out coverage.
Working with it
The ontology is loaded and queried from Python with Owlready2, with HermiT or Pellet as optional reasoners. A set of SPARQL queries answers questions about insurance and services. Validation scripts check the ontology's structure and run its competency questions: the questions an ontology must be able to answer, written down before it is built.
Report
The project includes a full written report in LaTeX, with architecture and ontology diagrams drawn in TikZ and its own bibliography.
See also
Pictures
