Named entity recognition (NER) finds the names of real-world things in text, such as people, companies, places and products, and labels what kind each one is.
In plain words
NER reads a sentence and marks spans of words that name something, then tags each one: this is a person, this is an organisation, this is a place, this is a date. It is a token classification task, because the model labels each word or piece of a word.
A quick example
spaCy's docs use a sentence about Sebastian Thrun starting work on self-driving cars at Google in 2007. Its model tags Sebastian Thrun as a person, Google as an organisation and 2007 as a date. An app could then link that text to the people and companies it mentions.
Where you'll meet it
- Extracting facts from news, contracts and support tickets.
- Search, analytics and data cleaning pipelines.
Related
Sources
- DocsLinguistic Features: Named Entities, spaCy documentation · read 29 Sept 2026
- DocsToken classification, Hugging Face Transformers documentation · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.