yago3_10 — YAGO3 knowledge graph benchmark (YAGO3-10, 2018)

Description

A subset of the YAGO3 knowledge base, extracted from Wikipedia, WordNet and GeoNames, that is used as a benchmark for knowledge graph completion (link prediction): the YAGO3-10 dataset of Dettmers et al. (2018), with the entities that have at least 10 relations each. Each node is an entity, mostly people, places, organizations and works, named as in YAGO (after the titles of their Wikipedia articles), and a directed edge from entity A to entity B is a triple (A, relation, B) of one of the 37 relations, e.g. that a person playsFor a team, isCitizenOf a country, or hasGender a gender. Pairs of entities may be connected by several edges with different relations.

The triples are divided into training, validation and test sets (5,000 triples each for validation and testing) by the benchmark; all of them are included, and the set is given as an edge property.

The network has the following properties:

(Node) name: the name of the entity in YAGO, e.g. Chatou or Boo_Young-tae.
(Edge) relation: the YAGO relation, e.g. isLocatedIn.
(Edge) split: the set of the benchmark the triple belongs to: train, valid or test.

Tags

Informational Relatedness Unweighted Multigraph Metadata

Added

2026-10-27

Citation

T. Dettmers, P. Minervini, P. Stenetorp, S. Riedel, "Convolutional 2D Knowledge Graph Embeddings." Proceedings of the AAAI Conference on Artificial Intelligence 32(1), 1811-1818 (2018) 10.1609/aaai.v32i1.11573 Sci-Hub
F. Mahdisoltani, J. Biega, F. M. Suchanek, "YAGO3: A Knowledge Base from Multilingual Wikipedias." 7th Biennial Conference on Innovative Data Systems Research (CIDR) (2015) www.cidrdb.org

Upstream URL OK

https://github.com/TimDettmers/ConvE

Upstream license

CC BY 3.0 (YAGO3); MIT (ConvE repository), https://creativecommons.org/licenses/by/3.0/

Networks

Tip: hover over a column header for a description.

Column legend

$\left<k\right>$: Mean degree
$\sigma_k$: Standard deviation of the degrees
$\lambda_h$: Largest eigenvalue of the Hashimoto (non-backtracking) matrix
$\tau$: Random walk mixing time
$r$: Degree assortativity
$c$: Global clustering coefficient
$\oslash$: (Pseudo-) diameter
$S$: Relative size of the largest component
NPs: Node properties
EPs: Edge properties

Name	Nodes	Edges	$\left<k\right>$	$\sigma_k$	$\lambda_h$	$\tau$	$r$	$c$	$\oslash$	$S$	Kind	Mode	NPs	EPs	gt	GraphML	GML	csv
yago3_10	123,182	1,089,040	8.84	185.46	101.07	41.50	-0.04	0.00	11	1.00	Directed	Unipartite	name	relation split	5.2 MiB	9.4 MiB	9.2 MiB	7.9 MiB

Ridiculograms^*

^* These are automatically generated force-directed visualizations, and can be quite meaningless for networks both big and small. They should not be taken seriously as sources of scientific insight. See here for a discussion.