Natural language database querying

• US 20090112835A1
• Filed: 12/20/2007
• Published: 04/30/2009
• Est. Priority Date: 10/24/2007
• Status: Abandoned Application

• Current Assignee: Marvin Elder
• Original Assignee: Marvin Elder
• Inventors: Elder, Marvin
• Application Number:
  US12/004,255
• Publication Number:
  US 20090112835A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: F16K 15/16 with tongue-shaped laminae

Abstract

The invention is natural language database querying (NLDQ), by a) allowing a user to type an information request using phrases stated in a natural language such as English, b) matching the phrases in the request to semantic objects stored in a hierarchy of ontologies, c) automatically converting the matched semantic phrases into the names of synchronized conceptual objects stored in the same taxonomy, d) automatically mapping the conceptual objects to metadata objects in one or more target databases; e) from the mapped metadata objects, generating a database query stated in a database language such as Structured Query Language (SQL), and f) executing the database query against the target database(s).

Claims

1. A method, comprising sequentially:

receiving a natural language request, the natural language request being composed of phrases stated in, or convertible to, the common vernacular used in every-day natural language;

parsing phrases in the natural language request against semantic phrases stored in a hierarchical set of ontologies;

mapping the semantic phrases to conceptual objects that have been synchronized with semantic phrases in the same taxonomy;

transforming the mapped conceptual objects, through a rules-based inferencing process, to database objects stored in a repository of database metadata for one or more target databases previously prepared through mappings to ontology objects;

from the mapped database objects, generating an executable database query statement in a suitable database query language; and

sending the database query results to the requesting user.

2. The method of claim 1 further comprising discovering a new phrase entered by the user in a natural language, and upon clarification by the user, adding this new phrase to an existing semantic phrase repository, and adding the phrase in a base language in a synchronized concept model repository, the two synchronized sets of objects defining a ‘

semantified’

ontology taxonomy.

3. The method of claim 1 further comprising generating a join clause of an executable database query statement by automating a navigation algorithm that connects all mapped database table objects that were identified indirectly in a single user request.

4. The method of claim three wherein the connections form an optimum path between table relationships in single or federated target databases.