Relationship Types

RelationshipType Entity1 Type Entity2 Type Relationships Covered
AffiliatedWith Person Organization Indicates any professional or academic relationship between the Person and the Organization entities.

The relationship can be any of these or other similar relationships:

  • Person is studying in or studied in Organization
  • Person is working with or worked with Organization
  • Person has been offered a job at Organization
Note: This is an indicative list of the relationships this type covers.

For example,

James has studied from the University of Toronto and works at ABC Corp.

Here, two relationships can be parsed:

Entity1 = James, RelationshipType = AffiliatedWith, Entity2 = University of Toronto

Entity1 = James, RelationshipType = AffiliatedWith, Entity2 = ABC Corp

LivesIn Person Location Indicates a relationship between the Person and the Location entities.
The relationship can be any of these:
  • Person stays in or stayed in Location
  • Person shifted to Location
  • Person was born in Location
  • Person died at Location
Note: This is an indicative list of the relationships this type covers.

For example,

John Jamison, a National Weather Service meteorologist in Galveston, reported that a massive hurricane was about to hit the East Coast the next day.

Entity1 = John Jamison, RelationshipType = LivesIn, Entity2 = Galveston
OrgBasedIn Organization Location Indicates that the Organization has at least one of its offices in the Location.

The Location can be a branch office, development office, headquarters, and the like.

For example,

HSBC Holdings Plc. is headquartered in London, United Kingdom.

Here, two relationships can be parsed:

Entity1 = HSBC Holdings Plc., RelationshipType = OrgBasedIn, Entity2 = London

Entity1 = HSBC Holdings Plc., RelationshipType = OrgBasedIn, Entity2 = United States of America

LocatedIn Location Location Indicates the relationship between two different locations, where one of the entities is geographically contained within the other entity.
Example 1

Canberra is the capital of Australia.

Here,

Entity1 = Canberra, RelationshipType = LocatedIn, Entity2 = Australia

Example 2

India has as its capital New Delhi.

Here,

Entity1 = India, RelationshipType = LocatedIn, Entity2 = New Delhi

Negative Person

Organization

Location

Organization

Location

Indicates that none of the above relationship types could be parsed between the two corresponding entities.

For example,

New Delhi and New York are good places to live in.

On parsing this input text, none of the supported relationship types are parsed between any pair of identified entities. Hence it can be broken down into Negative relationship types between the identified entities:

Entity1 = New Delhi, RelationshipType = Negative, Entity2 = New York

Note: You can connect a Splitter stage at the output of the Relationship Extractor stage to extract the identified relationship types and the corresponding pair of entities joined by the relationship. The splitter stage converts the hierarchal output of this stage to a flat output.

Example

In case of a complex input text, multiple possible relationship type combinations might be parsed for the same sentence.

For example,

James McCarthy has settled in New York, United States as director of ABC Technologies.

When the Relationship Extractor stage parses this input text using the relationship types selected in the stage options, the relationships found are:

Relationship 1
Entity1 = James McCarthy, Entity1 Type = Person, RelationshipType = LivesIn, Entity2 = New York, Entity2 Type = Location
Relationship 2
Entity1 = James McCarthy, Entity1 Type = Person, RelationshipType = AffiliatedWith, Entity2 = ABC Technologies, Entity2 Type = Organization
Relationship 3
Entity1 = ABC Technologies, Entity1 Type = Organization, RelationshipType = OrgBasedIn, Entity2 = United States, Entity2 Type = Location
Relationship 4
Entity1 = ABC Technologies, Entity1 Type = Organization, RelationshipType = OrgBasedIn, Entity2 = New York, Entity2 Type = Location
Relationship 5
Entity1 = James McCarthy, Entity1 Type = Person, RelationshipType = LivesIn, Entity2 = United States, Entity2 Type = Location
Relationship 6
Entity1 = New York, Entity1 Type = Location, RelationshipType = LocatedIn, Entity2 = United States, Entity2 Type = Location