Support Center

Relationship Detection

Last Updated: Oct 03, 2013 10:56AM EDT

The relationship detection identifies connections between named entities. This feature is only available via the API.
Relationship detection between entities is only available in English and is not currently available in Semantria’s other supported languages (French, Spanish, Portuguese, and German). 

Semantria is able to identify many relationships right out-of-the-box, including:

  • Career is the relationship between PersonCompany and Job Title entities.
  • Employment is the relationship between Company and Person entities.
  • Former career is the same as the Career relationship, but for former activities.
  • Former Employment is the same as the Employment relationship, but for former activities.
  • Occupation is the relationship between Person and Job Title entities.
  • Former occupation is the same as the Occupation relationship but for former activities.
  • Headquarters is the relationship between Company and Place entities.
  • Quotation is the relationship between Person and Quote entities.

***Semantria does not currently support Competitor relationships.
Consider the following excerpt from a New York Times article, analyzed by Semantria:
“The law’s a train wreck,” Speaker John A. Boehner said of the health care law, the Affordable Care Act, on Wednesday morning. “It’s time to protect American families from this unworkable law.”
Through the API, Semantria will extract the Quotation relationship:
            "relations": [{
                        "type": "named",
                        "extra": "said",
                        "entities": [{
                                    "title": "Speaker John A. Boehner",
                                    "entity_type": "Person"
                                    "title": "“The law’s a train wreck,”;“It’s time to protect American families from this unworkable law.”",
                                    "entity_type": "Quote"
                        "relation_type": "Quotation",
                        "confidence_score": 1.0
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found