From controlled vocabulary to queryable knowledge graph — how ArcaQ traverses all 6 levels of semantic maturity
A curated, deduplicated set of preferred terms and labels that eliminates ambiguity across enterprise datasets. The mandatory foundation before any semantic layer can be built.
ArcaQ Glossary Agent — Automatically extracts and normalises terms from structured and unstructured sources. Preferred labels, alternate labels, and scope notes stored as RDF skos:Concept with full provenance and audit trail.
Native
SKOS RDF
ISO 25964
Formal schemas that validate structure, lineage, and discoverability. Bridges raw data assets to semantic meaning through internationally recognised W3C standards.
ArcaQ Connect Agent — Applies Dublin Core, Schema.org JSON-LD, and PROV-O provenance triples to every ingested asset. DCAT-compliant catalog entries generated automatically via the Catalog Agent.
Native
Schema.org
PROV-O
Hierarchical organisation of concepts using broader/narrower relationships (SKOS). Enables drill-down SPARQL queries, faceted search, and automatic category inheritance across the graph.
ArcaQ Semantic Builder Agent — Auto-generates domain taxonomies from ingested data using LLM-powered classification. OWL rdfs:subClassOf hierarchies propagate classifications automatically via SPARQL CONSTRUCT rules.
Native
SKOS Hierarchical
OWL rdfs:subClassOf
Associative and equivalence relationships between concepts across domains. Enables synonym-aware querying, cross-domain federation, and automatic terminology alignment.
ArcaQ Knowledge Graph — Stores skos:related, skos:exactMatch, and skos:closeMatch edges natively. The Brain Agent leverages these for synonym-transparent SPARQL federation across siloed enterprise datasets.
Native
SKOS Associative
Cross-domain
Formal logic layer adding classes, properties, axioms, and constraints. Enables automated reasoning, inference, and compliance enforcement. The layer protected by ArcaQ's 26 patent claims.
ArcaQ OWL 2 Ontology — SHACL shapes enforce data quality. Apache Jena OWL reasoner derives implicit knowledge automatically. The Shield Agent maps GDPR, ISO 42001, and sector regulations directly as OWL restrictions and SPARQL compliance rules.
Core Layer
26 Patent Claims
OWL 2 + SHACL
Queryable, living synthesis of all upstream layers. Integrates structured data (entity resolution), unstructured content (NLP/NER), and domain thesauri via entity linking into a unified sovereign graph.
ArcaQ Sovereign Knowledge Graph — Apache Jena TDB2 triple store. Bi-temporal versioning tracks knowledge evolution. CAG Architecture delivers 99.9% answer precision from certified graph facts. Full SPARQL endpoint on-premise.
Target Layer
Bi-temporal RDF
99.9% CAG Precision
Structured sources (SQL, Excel, APIs) enter via entity resolution → normalised to controlled vocabulary → catalogued with metadata standards → classified into the taxonomy hierarchy.
Unstructured sources (PDFs, emails, reports) are processed by NLP/NER → entities linked to the thesaurus layer → promoted to OWL class instances → stored as RDF triples in the Knowledge Graph.
The result: a sovereign, fully queryable Knowledge Graph that unifies all 6 ontology maturity levels — 100% on-premise, zero cloud dependency.