Skip to content

Named Entity Standardization

Named entity standardization (NES) is the process of identifying the preferred ontology (or database) term for a concept from many equivalent concepts. This process typically relies on the existence of high quality semantic mappings such as equivalence relationships and some usage of inference tools.

The algorithm for NES is:

  1. Generate equivalence classes, e.g. using semantic mappings from ontologies, databases, and Biomappings
  2. Apply inference and reasoning, e.g., such as chaining rules
  3. Apply priority rules (often, a ranked list in terms of prefixes) to identify the preferred term for each equivalent class
  4. Generate a surjective mapping (i.e., means many keys can point to the same value) that can be used to map any given entity to its preferred entity
  5. Apply this mapping to data, such as the results from named entity recognition (NER) and named entity normalization (NEN).

NES Tools