wikidata5m — Wikidata knowledge graph (Wikidata5M, 2019)

Description

A knowledge graph of the entities of Wikidata that have a page in the English Wikipedia, from the Wikidata5M dataset, made for the KEPLER model of knowledge embedding and language representation, and used as a benchmark for link prediction. Each node is an entity of Wikidata (e.g. a person, a place, a work or a concept) that has a Wikipedia page, and a directed edge from entity A to entity B is a statement (A, relation, B) of Wikidata between two such entities, with one of 822 relations (Wikidata properties, e.g. P31, "instance of", or P27, "country of citizenship"). The data come from the Wikidata dump of 2019-07-31 and the English Wikipedia dump of 2019-08-01; the entities without a Wikipedia page, and those with few statements, were left out by the authors.

The triples are given in the "transductive" split of the benchmark (the same entities in the training, validation and test sets), and all of them are included here, with the set as an edge property. The Wikipedia texts of the entities, also part of the dataset, are not included; the "inductive" split, with the same triples split by entity, is also left out. A larger knowledge graph of Wikidata of 2015 is in the entry ogbl_wikikg2, and the complete Wikidata in wikidata.

The network has the following properties:

(Node) wikidata_id: the Wikidata identifier of the entity, e.g. Q42.
(Node) name: the name of the entity (its first alias in the dataset, usually its English label).
(Edge) relation: the Wikidata property of the relation, e.g. P31.
(Edge) relation_name: the name of the relation, e.g. instance of.
(Edge) split: the set of the benchmark the triple belongs to: train, valid or test.

Tags

Informational Relatedness Unweighted Multigraph Metadata

Added

2026-10-28

Citation

X. Wang, T. Gao, Z. Zhu, Z. Zhang, Z. Liu, J. Li, J. Tang, "KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation." Transactions of the Association for Computational Linguistics 9, 176-194 (2021) 10.1162/tacl_a_00360 Sci-Hub

Upstream URL OK

https://deepgraphlearning.github.io/project/wikidata5m
https://doi.org/10.5281/zenodo.5546383

Upstream license

CC BY 4.0, https://creativecommons.org/licenses/by/4.0/

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
wikidata5m	4,594,485	20,624,575	4.49	892.45	788.28	3044.39	-0.08	0.00	104	1.00	Directed	Unipartite	wikidata_id name	relation relation_name split	116.9 MiB	199.7 MiB	193.7 MiB	190.8 MiB