Suggested Upper Merged Ontology (SUMO)

The Suggested Upper Merged Ontology (SUMO) is one of the largest and most widely used formal upper ontologies, providing a comprehensive framework for representing general concepts and relationships across all domains of knowledge. SUMO and its domain ontologies are used for research and applications in search, linguistics, automated reasoning, and artificial intelligence. SUMO is unique in being formally mapped to the entire WordNet lexicon, enabling semantic integration between natural language and formal knowledge representation. The ontology covers abstract and concrete entities, processes, attributes, relations, and events, supporting logical inference and knowledge discovery. SUMO is open source and maintained by an active community, with ongoing extensions and domain-specific modules for specialized applications. By providing a rigorous semantic foundation, SUMO facilitates interoperability, data integration, and advanced reasoning in knowledge-based systems.

Example Usage: Use SUMO as the upper ontology for a knowledge graph, mapping domain-specific concepts (e.g., “vehicle,” “disease,” “financial transaction”) to SUMO classes and leveraging its logical axioms for automated reasoning and semantic search.

Metrics & Statistics

Graph Statistics

Total Nodes

288016

Total Edges

496645

Root Nodes

77015

Leaf Nodes

197102
Knowledge Coverage Statistics

Classes

4525

Individuals

80034

Properties

587
Hierarchical Metrics

Maximum Depth

9

Minimum Depth

0

Average Depth

1.04

Depth Variance

1.39
Breadth Metrics

Maximum Breadth

77015

Minimum Breadth

10

Average Breadth

19045.20

Breadth Variance

739917637.16
LLMs4OL Dataset Statistics

Term Types

80280

Taxonomic Relations

7174

Non-taxonomic Relations

310

Average Terms per Type

165.53

Usage Example

Use the following code to import this ontology programmatically:

from ontolearner.ontology import SUMO

ontology = SUMO()
ontology.load("path/to/SUMO-ontology.owl")

# Extract datasets
data = ontology.extract()

# Access specific relations
term_types = data.term_typings
taxonomic_relations = data.type_taxonomies
non_taxonomic_relations = data.type_non_taxonomic_relations