QUESTION TYPE AND DOMAIN IDENTIFYING APPARATUS AND METHOD

• US 20120101807A1
• Filed: 10/25/2011
• Published: 04/26/2012
• Est. Priority Date: 10/25/2010
• Status: Active Grant

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: HEO, Jeong, Jang, Myung Gil, Kim, HyunKi, Ryu, Pum Mo, Lee, Changki, Lee, Chung Hee, Lim, Soojong, Choi, Miran, Yoon, Yeo Chan, Oh, Hyo-Jung, Park, Sang Kyu, Shin, Ji Ae, Choi, Yoonjae
• Granted Patent: US 8,744,837 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 16/3331 Query processing

Abstract

A question type and domain identifying apparatus includes: a question type identifier for recognizing the number of words of a user'"'"'s question to identify whether the user'"'"'s question is a query for performing information searching or a question for performing a question and answer (Q&A); a question domain distributor for distributing one of plural preset domain specialized Q&A engines, as a Q&A engine of the user'"'"'s question based on the recognized word number; and a Q&A engine block, including the domain specialized Q&A engines, for selectively performing information searching or a Q&A with respect to the user'"'"'s question in response to the distribution of the question domain distributor.

Claims

1. A question type and domain identifying apparatus, the apparatus comprising:

a question type identifier for recognizing the number of words of a user'"'"'s question to identify whether or not the user'"'"'s question is a query for performing information searching or a question for performing a question and answer;

a question domain distributor for distributing one of plural preset domain specialized question and answer engines, as a question and answer engine of the user'"'"'s question based on the recognized word number; and

a question and answer engine block, including the domain specialized question and answer engines, for selectively performing information searching or a question and answer with respect to the user'"'"'s question in response to the distribution of the question domain distributor.

2. The apparatus of claim 1, wherein the question type identifier includes:

a word number recognition unit for classifying the user'"'"'s question as the query for information searching when the recognized word number is one;

a declinable word/postposition identification unit for classifying the user'"'"'s question as a natural language question when the recognized word number is two or more and a predicate or postposition is determined to be included in the user'"'"'s question; and

a question and answer type recognition unit for classifying the user'"'"'s question as the query or a keyword array type question based on the question and answer type identification result when a predicate or postposition is not included in the user'"'"'s question.

3. The apparatus of claim 2, wherein when a weight value of the question and answer type does not exceed a preset threshold value, the question and answer type recognition unit classifies the user'"'"'s question as the query, and when the weight value of the question and answer type exceeds the preset threshold value, the question and answer type recognition unit classifies the user'"'"'s question as the keyword array type question.

4. The apparatus of claim 1, wherein the question domain distributor distributes a domain with respect to the user'"'"'s question based on a law dictionary previously established for each domain or a learning model established by using learning data.

5. The apparatus of claim 4, wherein the question domain distributor includes:

a language analyzer for analyzing the language of the user'"'"'s question;

a law-based domain distribution block for identifying and distributing a domain for the user'"'"'s question based on the law dictionary though the language analysis result;

a learning model database for storing the learning model established by analyzing the language with respect to the learning data and through mechanical leaning of learning quality; and

a learning model-based domain distribution block for distributing a domain with respect to the user'"'"'s question based on the leaning model when a domain is not distributed based on the law dictionary.

6. The apparatus of claim 5, wherein the learning model is established by using leaning data with respect to representative questions of respective domains.

7. The apparatus of claim 1, further comprising:

a contents search filtering engine for restricting a search target document in the distributed question and answer engine based on question subject information and area information identified through the analysis of the user'"'"'s question. 25

8. The apparatus of claim 7, wherein the contents search filtering engine includes:

a question subject filtering module for identifying a question subject based on a keyword included in the user'"'"'s question and restricting a range for document searching based on the identified question subject; and

an area filtering module for identifying an interested area keyword included in the user'"'"'s question and restricting the range for the document searching to an area related to the identified interested area keyword.

9. The apparatus of claim 8, wherein the question subject filtering module includes:

a language analyzer for analyzing the language of the user'"'"'s question to extract a keyword;

an index keyword-based subject filtering block for identifying first subject filtering information by using a statistics method utilizing frequency information of the extracted keyword based on a predefined index keyword dictionary;

a learning model database for storing the learning model established through analyzing the language with respect to the learning data and through mechanical leaning of learning quality;

a learning model-based subject filtering block for identifying second subject filtering information based on the learning model with respect to the learning quality extracted from the user'"'"'s question; and

a subject filtering integrating block for integrating the first and second subject filtering information to restrict the range for the document searching with respect to the identified question subject.

10. The apparatus of claim 8, wherein the area filtering module includes:

a language analyzer for analyzing the language of the user'"'"'s question to identify a named entity;

an interested area keyword identifier for selecting an interested area keyword target with respect to the identified named entity and identifying area information based on a predefined interested area keyword dictionary;

an area ambiguity elimination block for eliminating ambiguity of the identified area information by using other keywords related to the interested area keyword; and

a statistics-based representative area identification unit for restricting the range for document searching with respect to the related area based on the frequency of the interested area keyword and location information.

11. A question type and domain identifying method, the method comprising:

recognizing the number of words of a user'"'"'s question to identify whether or not the user'"'"'s question is a query for performing information searching or a question for performing a question and answer;

distributing one of plural preset domain specialized question and answer engines, as a question and answer engine of the user'"'"'s question, based on the recognized word number; and

selectively performing information searching or a question and answer with respect to the user'"'"'s question by using the domain specialized and question and answer engine distributed as the question and answer engine.

12. The method of claim 11, wherein said recognizing the number of words of a user'"'"'s question includes:

classifying the user'"'"'s question as the query for information searching when the recognized word number is one;

checking whether a predicate or postposition is included in the user'"'"'s question when the recognized word number is two or more;

classifying the user'"'"'s question as a natural language question when the predicate or postposition is included in the user'"'"'s question; and

classifying the user'"'"'s question as the query or a keyword array type question based on the question and answer type identification result when the predicate or postposition is not included in the user'"'"'s question.

13. The method of claim 12, further comprising:

classifying the user'"'"'s question as the query when a weight value of the question and answer type does not exceed a preset threshold value; and

classifying the user'"'"'s question as the keyword array type question when the weight value of the question and answer type exceeds the preset threshold value.

14. The method of claim 11, wherein said distributing one of plural preset domain specialized question and answer engines distributes a domain with respect to the user'"'"'s question based on a law dictionary previously established for each domain or a learning model established through analyzing the language of learning data and through mechanical leaning of learning quality.

15. The method of claim 14, wherein said distributing one of plural preset domain specialized question and answer engines includes:

analyzing the language of the user'"'"'s question;

identifying and distributing a domain for the user'"'"'s question based on the law dictionary though the language analysis result; and

distributing a domain with respect to the user'"'"'s question based on the leaning model when a domain is not distributed based on the law dictionary.

16. The method of claim 11, further comprising:

restricting a search target document in the distributed question and answer engine based on question subject information and area information identified through the analysis of the user'"'"'s question.

17. The method of claim 16, wherein said restricting a search target document includes:

identifying a question subject based on a keyword included in the user'"'"'s question and restricting a range for document searching based on the identified question subject; and

identifying an interested area keyword included in the user'"'"'s question and restricting the range for the document searching to an area related to the identified interested area keyword.

18. The method of claim 17, wherein said identifying a question subject includes:

analyzing the language of the user'"'"'s question to extract keywords;

identifying first subject filtering information by using a statistics method utilizing frequency information of the extracted keywords based on a predefined index keyword dictionary;

identifying second subject filtering information based on a learning model established through analyzing the language of learning data and through mechanical leaning of learning quality; and

integrating the first and second subject filtering information to restrict the range for the document searching with respect to the identified question subject.

19. The method of claim 17, wherein said identifying an interested area keyword includes:

analyzing the language of the user'"'"'s question to identify a named entity;

selecting an interested area keyword target with respect to the identified named entity and identifying area information based on a predefined interested area keyword dictionary;

eliminating ambiguity of the identified area information by using other keywords related to the interested area keyword; and

restricting the range for document searching with respect to the related area based on the frequency of the interested area keyword and location information.

20. The method of claim 19, wherein said restricting the range for document searching includes:

statistically identifying representative area information based on the frequency of the interested area keyword and location information; and

restricting the range for document searching to the identified representative area information.