Skip to main content
The Named Entities (NER) activity is designed to use Natural Language Processing (NLP) to extract named entities from unstructured documents, such as contracts, letters, orders, press releases, and other documents with no specific structure that can be described using rules. To process these documents using a Named Entities (NER) activity, you need to map the named entities to the skill fields into which the entity values will be extracted. This activity will then analyze the document and extract the named entities into their corresponding fields. You can also set up named entity extraction for fields extracted by other activities. Suppose you know that organization names and addresses that you need to extract are located in the first paragraph of each contract. You can extract the first paragraph using a Segmentation activity, and then extract company names and addresses from this paragraph using a Named Entities (NER) activity. This approach is more reliable than extracting named entities from the entire document, since you can control the specific area where those entities are extracted from.
The activity only supports fields of type Text that have data type set to Text, Date, or Money.

​
Setting up a Named Entities (NER) activity

To set up a Named Entities (NER) activity:
  1. On the Activities tab, add a Named Entities (NER) activity to the document processing flow.
  2. On the Activity Properties pane, use the Source drop-down list to select a source that the activity will use to extract named entities from—either the whole document or a single field extracted by another activity.
  3. In the Output field, select fields into which the named entities will be extracted. The output fields must be either on the same nesting level as the source field or one level below it.
  4. Click Create Mapping. In the dialog that will open, select which named entities will be extracted to each field in the Entity to extract list. Click Save. You can edit the mapping at any time by clicking Edit Mapping.
  5. Click Test Skill to test your skill and analyze the named entity extraction results on the Results tab.

​
Supported Named Entities

Entity nameDescriptionExampleSupported data typesSupported languages
PersonNames of peopleJohn Doe, Jane SmithTextEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
LocationNames of locationsAnytown, Corporate PlaceTextEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
OrganizationNames of organizationsABBYY, Acme Corp.TextEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
AddressAddresses123 Main Str., Anytown AB 45678, 950 Acacia Avenue 50, Anytown, AB 12345, USATextEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
MoneyAmounts of money$2670.00, 199 dollars 99 centsText, Amount of moneyEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
DateDatesNovember 14, 2009, 11/14/2009Text, DateEnglish, Russian, German, French, Spanish, Japanese, Italian, Portuguese (Standard), Dutch
DurationTime periodsTwelve (12) months, 4 daysTextEnglish, Russian, German, French, Spanish, Italian, Portuguese (Standard), Dutch