Dictionary creation device and dictionary creation method

• US 7,921,113 B2
• Filed: 08/02/2006
• Issued: 04/05/2011
• Est. Priority Date: 12/26/2003
• Status: Active Grant

• Current Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Kutsumi, Hiroshi, Naito, Eiichi, Ozawa, Jun, Yoshida, Hideyuki
• Primary Examiner(s): Abel-Jalil, Neveen
• Assistant Examiner(s): Syed, Farhan M
• Application Number:
  US11/497,357
• Publication Number:
  US 20060271527A1
• Time in Patent Office: 1,707 Days
• Field of Search: None
• US Class Current: 707/740
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: H04N 21/482 End-user interface for program selection broadcast systems using EPGs H04H60/72

Abstract

A dictionary creation device and dictionary creation method which optimally create and update a dictionary for classifying, searching, or extracting text information in accordance with a changes in content of text information groups. The dictionary creation device includes a keyword extraction unit that extracts a keyword from inputted text information; a keyword statistics unit that finds statistics regarding an appearance of the keyword; a keyword assessment value calculation unit that calculates an assessment value of the extracted keyword based on the statistics regarding the appearance of the keyword; a determination unit that determines whether or not to register or delete the keyword based on the calculated assessment value; a dictionary registration and deletion unit which registers or deletes the keyword in or from a dictionary database based on a result of the determination performed by the determination unit; and the dictionary database.

Claims

1. A dictionary creation device that creates a dictionary which is used for searching, classifying, or filtering information written as text and in which keywords are registered per category, the dictionary creation device comprising:

a classification information acquisition unit that acquires classification information regarding categories and text information from at least a first information source and a second information source which differ from an information source for information written as text and searched;

a keyword extraction unit that extracts a keyword from the acquired text information;

a dictionary registration and deletion unit that registers or deletes the extracted keyword in dictionaries corresponding to the first information source and the second information source, in accordance with a category of the first information source and a category of the second information source, respectively, based upon the classification information acquired by said classification information acquisition unit and the keyword extracted by said keyword extraction unit;

a keyword database that stores the extracted keyword, said keyword database being a non-transitory computer-readable storage medium; and

a dictionary combining and editing unit that edits the category of the first information source in the dictionary corresponding to the first information source and the category of the second information source in the dictionary corresponding to the second information source to create, as a category level structure of a combined dictionary, a new category level structure including the category of the first information source and the category of the second information source, based on a degree of overlap between characteristic keywords that are keywords characterizing classification information regarding the category of the first information source and characteristic keywords that are keywords characterizing classification information regarding the category of the second information source,

wherein said dictionary combining and editing unit (i) compares a first set, which is a set of characteristic keywords in a first category included in the first information source, with a second set, which is a set of characteristic keywords in a second category included in the second information source, and (ii) edits and combines the dictionaries corresponding to the first information source and the second information source such that the second category is placed in a lower level subordinate to the first category as an intersecting set of the first set and the second set is less common to the first set and more common to the second set.

2. The dictionary creation device according to claim 1,

wherein said dictionary combining and editing unit determines a corresponding relationship between each category of the dictionaries that correspond to each information source based on a keyword that corresponds to each category in the classification information of the dictionaries that correspond to each information source.

3. The dictionary creation device according to claim 2,

wherein said dictionary combining and editing unit determines a corresponding relationship between each category using the number of common elements found in the keywords registered per category of each information source.

4. The dictionary creation device according to claim 1,

wherein said classification information acquisition unit acquires the classification information regarding the category and the text information via the Internet, which is used as one of the information sources.

5. The dictionary creation device according to claim 4,

wherein the information written as text is information related to a TV program.

6. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 1;

a dictionary storage unit that stores a dictionary created by said dictionary creation device; and

a search unit that searches the EPG data using the dictionary and output a search result.

7. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 1;

a dictionary storage unit that stores a dictionary created by said dictionary creation device;

a document vector creation unit that creates a document vector that is a pair of a keyword and an weight from the EPG data using the dictionary;

a search condition acquisition unit that acquires a search condition;

a document evaluation value calculation unit that calculates an evaluation value for the EPG data based on the search condition and the document vector; and

a search result output unit that outputs EPG data selected in accordance with the calculated evaluation value or an assessment value.

10. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 2;

a dictionary storage unit that stores a dictionary created by said dictionary creation device; and

a search unit that searches the EPG data using the dictionary and output a search result.

11. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 3;

a dictionary storage unit that stores a dictionary created by said dictionary creation device; and

a search unit that searches the EPG data using the dictionary and output a search result.

12. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 4;

a dictionary storage unit that stores a dictionary created by said dictionary creation device; and

a search unit that searches the EPG data using the dictionary and output a search result.

13. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

the dictionary creation device according to claim 5;

a dictionary storage unit that stores a dictionary created by said dictionary creation device; and

a search unit that searches the EPG data using the dictionary and output a search result.

14. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

a dictionary creation device according to claim 2;

a dictionary storage unit that stores a dictionary created by said dictionary creation device;

a document vector creation unit that creates a document vector that is a pair of a keyword and an weight from the EPG data using the dictionary;

a search condition acquisition unit that acquires a search condition;

a document evaluation value calculation unit that calculates an evaluation value for the EPG data based on the search condition and the document vector; and

a search result output unit that outputs EPG data selected in accordance with the calculated evaluation value or an assessment value.

15. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

a dictionary creation device according to claim 3;

a dictionary storage unit that stores a dictionary created by said dictionary creation device;

a document vector creation unit that creates a document vector that is a pair of a keyword and an weight from the EPG data using the dictionary;

a search condition acquisition unit that acquires a search condition;

a document evaluation value calculation unit that calculates an evaluation value for the EPG data based on the search condition and the document vector; and

a search result output unit that outputs EPG data selected in accordance with the calculated evaluation value or an assessment value.

16. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

a dictionary creation device according to claim 4;

a dictionary storage unit that stores a dictionary created by said dictionary creation device;

a document vector creation unit that creates a document vector that is a pair of a keyword and an weight from the EPG data using the dictionary;

a search condition acquisition unit that acquires a search condition;

a document evaluation value calculation unit that calculates an evaluation value for the EPG data based on the search condition and the document vector; and

a search result output unit that outputs EPG data selected in accordance with the calculated evaluation value or an assessment value.

17. An Electric Program Guide (EPG) TV show search system which searches EPG data, the system comprising:

a dictionary creation device according to claim 5;

a dictionary storage unit that stores a dictionary created by said dictionary creation device;

a document vector creation unit that creates a document vector that is a pair of a keyword and an weight from the EPG data using the dictionary;

a search condition acquisition unit that acquires a search condition;

a document evaluation value calculation unit that calculates an evaluation value for the EPG data based on the search condition and the document vector; and

a search result output unit that outputs EPG data selected in accordance with the calculated evaluation value or an assessment value.

8. A dictionary creation method for creating a dictionary which is used for searching, classifying, or filtering information written as text and in which keywords are registered per category, the dictionary creation method comprising:

acquiring classification information regarding categories and text information from at least a first information source and a second information source which differ from an information source for information written as text and searched;

extracting a keyword from the acquired text information;

registering or deleting the extracted keyword in dictionaries corresponding to the first information source and the second information source, in accordance with a category of the first information source and a category of the second information source, respectively, based upon the classification information acquired and the keyword extracted; and

editing the category of the first information source in the dictionary corresponding to the first information source and the category of the second information source in the dictionary corresponding to the second information source to create, as a category level structure of a combined dictionary, a new category level structure including the category of the first information source and the category of the second information source, based on a degree of overlap between characteristic keywords that are keywords characterizing classification information regarding the category of the first information source and characteristic keywords that are keywords characterizing classification information regarding the category of the second information source,

wherein said editing (i) compares a first set, which is a set of characteristic keywords in a first category included in the first information source, with a second set, which is a set of characteristic keywords in a second category included in the second information source, and (ii) edits and combines the dictionaries corresponding to the first information source and the second information source such that the second category is placed in a lower level subordinate to the first category as an intersecting set of the first set and the second set is less common to the first set and more common to the second set.

9. A program recorded on a non-transitory computer-readable storage medium for creating a dictionary which is used for searching, classifying, or filtering information written as text and in which keywords are registered per category, the program causing a dictionary creation device to perform steps comprising:

acquiring classification information regarding categories and text information from at least a first information source and a second information source which differ from an information source for information written as text and searched;

extracting a keyword from the acquired text information;

registering or deleting the extracted keyword in dictionaries corresponding to the first information source and the second information source, in accordance with a category of the first information source and a category of the second information source, respectively, based upon the classification information acquired and the keyword extracted; and

editing the category of the first information source in the dictionary corresponding to the first information source and the category of the second information source in the dictionary corresponding to the second information source to create, as a category level structure of a combined dictionary, a new category level structure including the category of the first information source and the category of the second information source, based on a degree of overlap between characteristic keywords that are keywords characterizing classification information regarding the category of the first information source and characteristic keywords that are keywords characterizing classification information regarding the category of the second information source,

wherein said editing (i) compares a first set, which is a set of characteristic keywords in a first category included in the first information source, with a second set, which is a set of characteristic keywords in a second category included in the second information source, and (ii) edits and combines the dictionaries corresponding to the first information source and the second information source such that the second category is placed in a lower level subordinate to the first category as an intersecting set of the first set and the second set is less common to the first set and more common to the second set.

18. A dictionary creation method for creating a dictionary which is used for searching, classifying, or filtering information written as text and in which keywords are registered per category, the dictionary creation method comprising:

acquiring classification information regarding categories and text information from at least a first information source and a second information source which differ from an information source for information written as text and searched

extracting a keyword from the acquired text information;

registering or deleting the extracted keyword in dictionaries corresponding to the first information source and the second information source, in accordance with a category of the first information source and a category of the second information source, respectively, based upon the classification information acquired and the keyword extracted;

storing, in a keyword database that is a non-transitory computer-readable storage medium, the extracted keyword; and

editing the category of the first information source in the dictionary corresponding to the first information source and the category of the second information source in the dictionary corresponding to the second information source to create, as a category level structure of a combined dictionary, a new category level structure including the category of the first information source and the category of the second information source, based on a degree of overlap between characteristic keywords that are keywords characterizing classification information regarding the category of the first information source and characteristic keywords that are keywords characterizing classification information regarding the category of the second information source,

wherein said editing (i) compares a first set, which is a set of characteristic keywords in a first category included in the first information source, with a second set, which is a set of characteristic keywords in a second category included in the second information source, and (ii) edits and combines the dictionaries corresponding to the first information source and the second information source such that the second category is placed in a lower level subordinate to the first category as an intersecting set of the first set and the second set is less common to the first set and more common to the second set.

19. A program recorded on a non-transitory computer-readable storage medium for creating a dictionary which is used for searching, classifying, or filtering information written as text and in which keywords are registered per category, the program causing a dictionary creation device to perform steps comprising:

acquiring classification information regarding categories and text information from at least a first information source and a second information source which differ from an information source for information written as text and searched;

extracting a keyword from the acquired text information;

registering or deleting the extracted keyword in dictionaries corresponding to the first information source and the second information source, in accordance with a category of the first information source and a category of the second information source, respectively, based upon the classification information acquired and the keyword extracted;

storing the extracted keyword in a keyword database; and

editing the category of the first information source in the dictionary corresponding to the first information source and the category of the second information source in the dictionary corresponding to the second information source to create, as a category level structure of a combined dictionary, a new category level structure including the category of the first information source and the category of the second information source, based on a degree of overlap between characteristic keywords that are keywords characterizing classification information regarding the category of the first information source and characteristic keywords that are keywords characterizing classification information regarding the category of the second information source

wherein said editing (i) compares a first set, which is a set of characteristic keywords in a first category included in the first information source, with a second set, which is a set of characteristic keywords in a second category included in the second information source, and (ii) edits and combines the dictionaries corresponding to the first information source and the second information source such that the second category is placed in a lower level subordinate to the first category as an intersecting set of the first set and the second set is less common to the first set and more common to the second set.