Named entity recognition
What is named entity recognition?
Named entity recognition (NER) is the process of detecting and categorizing anything that has a distinct and self-contained existence within a piece of text. It involves tagging words or phrases with their semantic meaning, such as individual, organization, product, location, date and more.
The process of identifying and classifying named entities in text can be implemented in several ways, including:
- The dictionary-based method is considered to be the simplest. It involves using a set vocabulary to find matching entities within the text.
- The rule-based method involves a predefined set of pattern- and context-based rules that are used for information extraction.
- The machine-learning method can be used to recognize an existing entity even if it has a minor spelling variation.
- The deep-learning method is considered to be the most accurate because it is able to correlate the semantic and syntactic relationship between various entities.
Benefits of named entity recognition
NER adds structure to previously unstructured text, allowing machine-learning algorithms to identify mentions of certain entities within a text. It is one of the building blocks of natural language understanding, a subset of natural language processing (NLP) that deals with helping machines to recognize the intended meaning of language.
The processes that NER enables, such as automating text summarization, locating mentions of particular entities across multiple documents and identifying the most important entities in a text, have a wide variety of business use cases. These include improving search and product recommendations, automating customer support processes, enabling chatbots to interact with humans and more.