Extract Named Entities

Getting started

With reciTAL's platform, you can detect several types of entities:

"person_name",
    "email_address",
    "phone_number",
    "social_security_number",
    "organisation",
    "localisation",
    "address",
    "url",
    "date",
    "date",
    "numeric_value",
    "currency_value",
    "currency_value",
    "percentage_value",
    "percentage_value"

To extract the entities of a document, simply follow the recipe:

Value extraction parameters

To refine the extraction of named entities, you can use the following parameters:

- types: search for a certain type of entities (eg: - monetary amounts)

value_type=ValueType.NER

- keywords: search for entities close to a specific keyword (e.g: search for "monetary amounts" around the keyword "Tax"). You can specificy the number "n" of words around the specified keyword.

keywords=[]
words_around=[]

Did this page help you?