Entity Linking — Named Entity Disambiguation
Entity Linking (Named Entity Disambiguation) · Also known as: named entity disambiguation, entity disambiguation, entity resolution to knowledge base, Varlık Bağlama (Entity Linking)
Entity linking is a natural-language-processing task that matches ambiguous entity mentions in text — people, places, organisations — to the correct record in a knowledge base such as Wikidata, DBpedia, or a domain dictionary. Surveyed and shaped by Milne and Witten (2008) and later neural approaches reviewed by Sevgili and colleagues (2022), it grounds free text into structured, unambiguous references used in knowledge-graph building and multi-source text analysis.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Entity linking fits when you have text containing named entities that are ambiguous and you can reach a target knowledge base to resolve them against. Named-entity recognition must already be complete, the target knowledge base must be accessible, and genuine entity ambiguity must be present for the task to add value. It is used to build knowledge graphs and to align entities across multiple text sources; a modest amount of text (around twenty documents or more) is enough to apply it meaningfully.
Strengths & limitations
- Grounds ambiguous text mentions into unique, machine-readable knowledge-base identifiers.
- Connects unstructured text to structured knowledge, enabling knowledge-graph construction and entity-centric search.
- Aligns the same real-world entity across multiple, independent text sources.
- Depends entirely on a prior named-entity-recognition step; errors there propagate into linking.
- Requires an accessible, reasonably complete knowledge base — missing entities cannot be linked.
- Resolving genuinely ambiguous mentions remains hard, especially for rare or newly emerging entities.
Frequently asked
How is entity linking different from named-entity recognition?
Named-entity recognition finds and labels the spans that are entities (this is a person, that is a place). Entity linking takes those recognised mentions and decides which specific real-world entity each one refers to, attaching a unique knowledge-base identifier. Linking assumes recognition is already done.
Which knowledge base should I link against?
Common choices are Wikidata and DBpedia for general entities, or a domain dictionary when your text covers a specialised field. The base must be accessible and cover the entities in your text; a base that lacks your domain's entities leaves many mentions unlinkable.
What happens to an entity that is not in the knowledge base?
It cannot be linked to a record. Such NIL mentions should be handled explicitly — marked as out-of-base rather than forced onto an incorrect candidate, which would inject errors into downstream analysis.
Why is disambiguation necessary if the name is already recognised?
A single name can denote many entities — the same surface form can point to different people, places, or organisations. Disambiguation uses the surrounding context to choose the single intended referent, which is the core of what entity linking adds beyond recognition.
Sources
- Milne, D. & Witten, I.H. (2008). Learning to Link with Wikipedia. CIKM (Proceedings of the 17th ACM Conference on Information and Knowledge Management). DOI: 10.1145/1458082.1458150 ↗
- Sevgili, O., Shelmanov, A., Arkhipov, M., Panchenko, A. & Biemann, C. (2022). Neural Entity Linking: A Survey of Models Based on Deep Learning. ACM Computing Surveys. DOI: 10.3233/SW-222986 ↗
How to cite this page
ScholarGate. (2026, June 1). Entity Linking (Named Entity Disambiguation). ScholarGate. https://scholargate.app/en/text-mining/entity-linking
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Named Entity RecognitionText mining↔ compare
- Relation ExtractionText mining↔ compare
- Scientific Text MiningText mining↔ compare