Named Entity Recognition: Definition and Examples

Named Entity Recognition (NER) is a fundamental subtask of natural language processing (NLP). It involves detecting mentions of real-world entities in raw text and categorizing them into predefined classes such as person names, organizations, geographical locations, dates, monetary amounts, or products.

This technology historically relied on rule-based and dictionary-based approaches, but modern methods predominantly use deep learning models, especially Transformer architectures like BERT or GPT. These models can understand the context surrounding each word to determine whether it is an entity and, if so, of which type.

In the context of prompt engineering, NER is particularly useful for structuring unstructured data. One can ask an LLM to extract named entities from a text, classify them by category, or use them as a basis for more complex tasks such as building knowledge graphs or targeted sentiment analysis.

NER has concrete applications in many fields: media monitoring (identifying actors mentioned in articles), legal analysis (extracting stakeholders from contracts), healthcare (spotting drug names and pathologies), and customer service (detecting products or services mentioned in complaints).