Ontology driven dictionary generation and ambiguity resolution for natural language processing

• US 9,372,924 B2
• Filed: 06/12/2012
• Issued: 06/21/2016
• Est. Priority Date: 06/12/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bostick, James E., Ganci, John M. Jr., Kaemmerer, John P., Trim, Craig M.
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US13/494,726
• Publication Number:
  US 20130332145A1
• Time in Patent Office: 1,470 Days
• Field of Search: 704/1, 704/9, 704/10, 706/47, 706/55
• US Class Current: 1/1
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 40/169 Annotation, e.g. comment data or footnotes
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A computer implemented method and system for natural language processing ambiguity resolution includes storing an ontology specifying a set of grammatical rules. A phrase comprising at least one current word to be processed is retrieved. A current word from the phrase is annotated with possible ontological classes according to the ontology. Any ontological rules associated with the possible ontological classes are retrieved. Ontological classes are eliminated based on the ontological rules. A surviving possible ontological class is determined to be an accurate ontological class for the current word. In another aspect of this disclosure, an ontology is stored in computer memory, the ontology having multiple ontological classifications, and word instances, each word instance associated with at least one of the ontological classifications. All word instances belonging to the selected ontological classification are retrieved.

Claims

1. A computer implemented method for natural language processing ambiguity resolution, comprising:

storing, in computer memory, an ontology specifying a set of grammatical rules, wherein the rules are associated with a word, with a sub-class of which the word belongs, and rules are associated with a greater class inclusive of the sub-class;

receiving, using a computer processor, a phrase having a natural language processing ambiguity and comprising at least one current word to be processed;

annotating, using the computer processor, each word from the phrase with an ontological class and sub-class to which the current word belongs according to the ontology, the subclass being associated with the class, the class further including at least one other sub-class, the class and sub-classes each having an associated one of the grammatical rules;

retrieving, using the processor, the grammatical rules associated with the ontological classes and sub-classes to which the current word belongs;

eliminating, using the processor, one or more ontological classes or sub-classes based on the grammatical rules associated with the respective classes and sub-classes;

determining, using the processor, a surviving ontological class and sub-class to be an accurate ontological class and sub-class for each word to resolve the natural language processing ambiguity;

storing, using the processor, the surviving ontological class and sub-class into an ontological library;

receiving, using the computer processor, a request to generate a dictionary for a selected ontological classification or sub-classification;

retrieving, using the computer processor, all word instances belonging to the selected ontological classification or sub-classification; and

compiling the retrieved word instances into a list to generate the requested dictionary for the selected ontological classification or sub-classification.

2. The method of claim 1, further comprising:

retrieving, using the processor, an ontological class or sub-class of an adjacent word from the phrase, the adjacent word being adjacent to the current word;

comparing, using the processor, the ontological class or sub-class of the adjacent word with a required ontological class or sub-class according to the ontological rule of one of the ontological classes or sub-classes to which the current word belongs; and

eliminating the one of the ontological classes or sub-classes to which the current word belongs if the ontological class or sub-class of the adjacent word does not match the required ontological class or sub-class.

3. The method of claim 2, further comprising:

repeating all steps for each word in the phrase until each word of the phrase is each annotated with only a single respective ontological class and sub-class from the ontology.

4. A system for natural language processing ambiguity resolution, comprising:

a computer processor; and

computer memory, the computer memory storing an ontology specifying a set of grammatical rules, wherein the rules are associated with a word, with a sub-class of which the word belongs, and rules are associated with a greater class inclusive of the sub-class, and a series of program instructions, wherein the program instructions are executable by the computer processor to;

receive, using the computer processor, a phrase having a natural language processing ambiguity and comprising at least one current word to be processed,

annotate, using the computer processor, each word from the phrase with an ontological class and sub-class to which the current word belongs according to the ontology, the sub-class being associated with the class, the class further including at least one other sub-class, the class and sub-classes each having an associated one of the grammatical rules,

retrieve, using the processor, the grammatical rules associated with the ontological classes and sub-classes to which the current word belongs,

eliminate, using the processor, one or more possible ontological classes or sub-classes based on the grammatical rules associated with the respective classes and sub-classes,

determine, using the processor, a surviving possible ontological class and sub-class to be an accurate ontological class and sub-class for each word to resolve the natural language processing ambiguity;

store, using the processor, the surviving ontological class and sub-class into an ontological library;

receive, using the computer processor, a request to generate a dictionary for a selected ontological classification or sub-classification;

retrieve, using the computer processor, all word instances belonging to the selected ontological classification or sub-classification; and

compile the retrieved word instances into a list to generate the requested dictionary for the selected ontological classification or sub-classification.

5. The system of claim 4, wherein the program instructions are further executable by the computer processor to:

retrieve, using the processor, an ontological class or sub-class of an adjacent word from the phrase, the adjacent word being adjacent to the current word;

compare, using the processor, the ontological class or sub-class of the adjacent word with a required ontological class or sub-class according to the ontological rule of one of the ontological classes or sub-classes to which the current word belongs; and

eliminate the one of the ontological classes or sub-classes to which the current word belongs if the ontological class or sub-class of the adjacent word does not match the required ontological class or sub-class.

6. The system of claim 5, wherein the program instructions are further executable by the computer processor to:

repeat all steps for each word in the phrase until each word of the phrase is each annotated with only a single respective ontological class and sub-class from the ontology.