Dx-Onto is an OWL core ontology representing digital-transformation project entities, relations and attributes in a knowledge graph. Tests on 10–1,000 synthetic projects and a heterogeneous Thai document corpus reported 85.2% domain fit versus 77.3% for COOT and enabled phase and strategic-dimension analyses absent from the baseline.
Key findings
- Dx-Onto scored 85.2% domain fit versus 77.3% for COOT and supported transformation-phase and strategic-dimension diagnostics not structurally available in the general baseline. Query scaling was demonstrated within the tested range, and a future hybrid LLM-ontology architecture is proposed.
Why this matters globally
Organizations often hold fragmented digital-project knowledge. An interpretable ontology can support semantic retrieval, cross-project comparison and a grounded layer for AI rather than relying only on keyword search or language models.
Thai researcher contribution
Prince of Songkla University and DEPA researchers used a real Thai document corpus, allowing Thai digital-transformation practice to shape a reusable semantic model.
Limitations to consider
Synthetic data do not capture production document complexity. Domain-fit scoring depends on definitions and raters not detailed in the abstract; one baseline cannot establish general superiority, and extraction accuracy, maintenance cost and an operational hybrid LLM system were not evaluated.
Verify the original sources
Applied System InnovationRead the original article↗DOI: 10.3390/asi9070146