Natural language query

• US 9,652,451 B2
• Filed: 05/08/2014
• Issued: 05/16/2017
• Est. Priority Date: 05/08/2014
• Status: Active Grant

• Current Assignee: Marvin Elder
• Original Assignee: Marvin Elder
• Inventors: Elder, Marvin
• Primary Examiner(s): BROWN, SHEREE N
• Application Number:
  US14/273,302
• Publication Number:
  US 20150324422A1
• Time in Patent Office: 1,104 Days
• Field of Search: 707722
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/86 Mapping to a database
• CPC Class Codes: G06F 16/907 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/908 using metadata automatically derived from the content
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30

Abstract

The invention allows end users to obtain immediate, accurate information from structured databases without writing complex database query commands. The invention allows two different, but synchronized, methods of end user information requests: Spoken or typed Natural Language requests, and a Visual Request Specification method. Furthermore, the invention provides a user the means of “teaching the system the correct interpretation” when an information request was misunderstood by the invention'"'"'s Natural Language Understanding module.

Claims

1. A computer-implemented method, comprising:

receiving, by a computer, a natural language request, said request comprises semantic phrases stated by voice or by text in plain English or other spoken or written language, wherein a plain meaning of a phrase consist of one word or contiguous words, and a term semantic asserting that a phrase has a different word meaning based on relevant context;

mapping, by a computer, a natural language request semantic phrases which are nouns to concept model objects stored in a hierarchical concept model repository part of a metadata repository database, resulting in a set of mapped concept model objects;

mapping, by a computer, said natural language request semantic phrases which are non-nouns against semantic phrases stored in a hierarchical semantic phrase repository part of the metadata repository database;

transforming, by a computer, mapped concept model objects into mapped database query objects by a natural language understanding process;

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

opening a virtual private network connection between the metadata repository database and a single or federated database associated with a current concept model;

executing by a computer, the generated executable database query statement;

capturing by a computer, database query results of the executed database query statement and sending said database query results to a requesting user;

maintaining said metadata repository database, consisting of a public domain open source OntoloNet metadata repository, consisting of a hierarchical concept model repository, a hierarchical subject area repository, and a hierarchical semantic phrase repository,

wherein said hierarchical concept model repository objects are mapped to subject area repository objects, said mapping is linked by identical object name;

said method further comprises a semantification process, comprising;

creating by a computer, a concept model object which includes a data model object of structural elements of a single or federated relational database consisting of tables and table columns, wherein said table comprises column width and type, indexes and table relationships,

wherein said concept model object components include adding by a computer, said created concept model object and all of said concept model object components to said hierarchical concept model repository portion of a metadata repository database;

creating by a computer, a subject area object and subject area object components, mapped by matching object name, to said concept model object and to said concept model object components;

adding by a computer, the created subject areas and subject area components to the hierarchical subject area repository portion of the meta data repository database repository;

converting by a computer, said mapped concept model object components, captured in a said natural language request, into visual requirement specification components;

mapping by a computer, said visual requirement specification components into database query components of the executable database query;

generating by a computer, a join clause of the executable database query statement by a navigation algorithm, executable by a computer processor, further comprising join clause components thereby forming an optimum path between the table relationships in the single or federated target databases;

discovering by a computer, a new non-noun phrase in a natural language request, and adding the discovered non-noun phrase to the semantic phrase repository portion of the metadata repository database repository;

replacing by a computer, visual request specification components of a failed user request user query of a user query database table by creating a user query, with a new set of semantically equivalent visual request specification components that map successfully to database query components of the executable database query, wherein replacement is characterized by a trained user, further comprising;

selecting by a computer, a failed user request user query from a database table of user query;

specifying by a computer, a new set of semantically equivalent visual request specification component;

employing a visual request specification process of selecting, using a visual drag-drop technique, database objects from a visible tree of entities and entity attributes, displaying said database objections from a visual request specification user interface page, to another section of the visual request specifications user interface page which contains selected visual request specification components, and setting visual request specification filter conditions, wherein each condition is composed of;

a selected Visual Request Specification component,

a relational operation from a list of relational operations terms consisting of one or more of the following;

equals, greater than, less than, is null, and is not null;

a user-specified value or another selected visual request specification component;

generating and executing by a computer processor, a database query with the replacement visual request specification components and filter conditions specified by the trained user;

storing the database query as a defined query;

creating by a computer, a user query request further comprising;

capturing by a computer, the visual request specification components of each natural language request as a user request in a user query database table, said user query request is one of the following;

successfully processed as the executable database query, or said user query request is failed;

and sending, by a computer, a code wherein said code indicates a reason for the failed user query request.