Automatic Clarification of Questions

The automatic clarification enhancements to English Query provide a way for a client application to automatically clarify certain classes of ambiguous questions when the user cannot or does not wish to provide the needed clarification. Although not all questions can be automatically clarified, these types of clarification questions can be automatically clarified for the user.

Proper Noun Clarification

Proper noun clarifications determine the entity to which an unknown proper noun refers, for example:

Question: How many awards did John Fortune win?

• A movie:

  SELECT * FROM movies WHERE name = 'John Fortune' 
  

• An actor:

  SELECT * FROM actors WHERE fname = 'John' and lname = 'Fortune'
  

Unique ID Clarification

Unique ID clarifications determine the entity to which an unknown ID refers, for example:

Question: Show the total sales for 423.

Clarification question: Which of these do you mean by "423"?

• A customer:

  SELECT * FROM customers WHERE id = 423
  

• An employee:

  SELECT * FROM employees WHERE id = '423'
  

Name/ID Clarification

Name/ID clarifications determine whether an unknown word refers to the name or ID of the specified entity, for example:

Question: What did customer Smith order?

Clarification question: Which of these do you mean by "Smith"?

• The customer whose name is Smith:

  SELECT * FROM customers WHERE name = 'Smith'
  

• The customer whose ID is Smith:

  SELECT * FROM customers WHERE id = 'SMITH'
  

Measurement Clarification

Measurement clarifications determines the value parameters used by an adjective of a specified entity, for example:

Question: List the old salespeople.

Clarification question: An old salesperson is a salesperson whose age is over ___

SELECT average(age) FROM salespeople

See Also

Adding Question Builder to an English Query Application

Testing an English Query Model