Object data retrieving device, object data retrieving method, and computer-readable recording medium containing recorder data

• US 6,912,523 B2
• Filed: 08/18/1998
• Issued: 06/28/2005
• Est. Priority Date: 08/18/1998
• Status: Expired due to Fees

• Current Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Original Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Inventors: Mitsugi, Tatsuya
• Primary Examiner(s): Ali, Mohammad
• Application Number:
  US09/424,661
• Publication Number:
  US 20020174084A1
• Time in Patent Office: 2,506 Days
• Field of Search: 707 1- 10, 707100-1041, 707200-205, 704/2, 704/8, 704/9, 701/201
• US Class Current: 1/1
• CPC Class Codes: G01C 21/3608 using speech input, e.g. using speech recognition
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

Object data is stored in association with a plurality of attribute words categorized according to sentence elements of a natural language, and an object data search is efficiently narrowed using a simple search criterion input. A database stores object data in association with a plurality of attribute words categorized according to the sentence elements of a natural language. An input unit receives an input of a search criterion in the form of a sentence of the natural language. A criterion retrieval unit analyzes the search criterion in the form of the sentence and retrieves one of a plurality of search words respectively corresponding to sentence element categories of the natural language. An object retrieval unit searches the database using each of the search words respectively associated with the sentence element categories, retrieves the object data associated with the attribute words that match a single search word or a plurality of search words wherein relationships existing between at least said one of said plurality of search words corresponding to the sentence element categories of the natural language are used with information features associated with the attribute words to further refine searching and an output apparatus for outputting the object data retrieved.

Claims

1. An object data search apparatus comprising:

a database for storing destination object data in association with a plurality of categorized attribute words categorized according to sentence elements of a natural language;

an input unit for receiving an input of search criteria in the form of a sentence of the natural language;

a criteria retrieval unit for analyzing the search criteria in the form of the sentence and retrieving one or a plurality of categorized search words respectively categorized corresponding to sentence element categories of the natural language;

a destination object retrieval unit for categorically searching sentence categories of the database using each of the categorized search words respectively associated with the sentence element categories, and retrieving the destination object data associated with the categorized attribute words that match a single search word or a plurality of search words in the same category, wherein filtering for attribute relation based on the grammatical structure of the natural language is performed; and

an output apparatus for outputting the destination object data thus retrieved.

2. The object data search apparatus according to claim 1, wherein said database stores destination object data at least associated with an attribute word having an agent of action category, an attribute word having an action category and an attribute word having an object of action category.

3. The object data search apparatus according to claim 1, said object retrieval unit retrieving a plurality of tuples and filtering the tuples so that overlapping tuples are filtered off.

4. The object data search apparatus according to claim 1, wherein said input unit is a voice input unit for receiving a voice input of search criteria in the form of a sentence of the natural language.

5. A destination-of-travel object data search method comprising the steps of:

receiving an input of search criteria in the form of a sentence of a natural language;

retrieving a plurality of search words from the search criteria input in the form of a sentence of a natural language by analyzing and categorizing the search criteria in accordance with a grammar of the natural language;

conducting a category-by-category search relative to a plurality of sentence element categories associated with a plurality of destination-of-travel object data items, based on the plurality of search words; and

retrieving the destination-of-travel object data associated with the attribute word that matches a single search word or a plurality of search words and outputting the destination-of-travel object data thus retrieved;

said conducting and retrieving using at least a search word having an agent of action category, a search word having an action category and a search word having an object of action category.

6. The destination-of-travel object data search method according to claim 5, said receiving step receiving a voice input of search criteria in the form of a sentence of the natural language.

7. A method of searching destination-of-travel object data comprising:

storing destination-of-travel object data in association with a plurality of categorized attribute words, wherein the attribute words are categorized and stored according to sentence elements of a natural language;

receiving an input of search criteria in the form of a sentence of a natural language;

analyzing the search criteria in the form of the sentence and retrieving at least one of a plurality of search words respectively corresponding to sentence element categories of the natural language;

searching sentence element categories of the database using each of the search words respectively associated with the sentence element categories, and retrieving the destination-of-travel object data associated with the categorized attribute words that match a single search word or a plurality of search words in the same category, wherein filtering for attribute relation based on the grammatical structure of the natural language is performed; and

outputting the destination-of-travel object data thus retrieved.

8. The method of searching destination-of-travel object data according to claim 7, said receiving step receiving a voice input of search criteria in the form of a sentence of the natural language.

9. A method for determining a destination based on a natural language query, comprising:

storing destination object data in association with categorized attribute words categorized according to sentence elements of the natural language, wherein the categories include agent-of-action, action, and object-of-action categories;

inputting a query utilizing a natural language sentence;

retrieving one or more categorized search words from the query such that each search word has an associated one of the categories corresponding to sentence elements of the natural language;

categorically searching the attribute words for a match with the retrieved search word and retrieving the destination object data associated with the attribute word that matches the search word,

wherein the category of attribute words searched by said searching step corresponds to the category of the search word, and

outputting the destination object data retrieved by said categorical search.

10. The method according to claim 9, said categorical searching including:

when the search word is in the agent-of-action category, searching the agent-of-action category for a match with the search word;

when the search word is in the action category, searching the action category for a match with the search word; and

when the search word is in the object-of-action category, searching the object-of-action category for a match with the search word.

11. The method according to claim 9, wherein the categories include agent-of-action, action, object-of-action, and key word categories.

12. The method according to claim 11, wherein the destination object data includes destination position information and name information of a destination.

13. The method according to claim 9, further comprising:

filtering for attribute relation based on a grammatical structure of the query.

14. The method according to claim 9, said inputting step inputting a voiced query utilizing the natural language sentence and said retrieving step retrieving one or more categorized search words from the voiced query.