DICTIONARY CREATION DEVICE

• US 20120310944A1
• Filed: 12/03/2010
• Published: 12/06/2012
• Est. Priority Date: 12/11/2009
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Mizuguchi, Hironori, Kusumura, Yukitaka, Kusui, Dai
• Granted Patent: US 9,600,468 B2
• Time in Patent Office: Days
• US Class Current: 707/740
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 40/242 Dictionaries

Abstract

A boundary word identification unit (103) identifies a boundary word belonging to a plurality of categories among words gathered in dictionary growth processing. Then, a category membership degree calculation unit (104) calculates, for each category to which the boundary word belongs, a category membership degree indicating a degree to which the boundary word belongs to the category on the basis of information recorded in a gathering process memory unit (108). Next, a category update unit (105) determines the category to which the boundary word belongs on the basis of the category membership degree calculated by the category membership degree calculation unit (104) and updates information stored in a gathered-by-category word memory unit (109) so that the determination result is reflected.

Claims

1. A dictionary creation device comprising:

an input/output process recording means for recording information indicating an input/output process for input words and output words output relating to said input words, in a dictionary growth process for gathering words in each category by repeatedly receiving input of words in each category, outputting from document data words related to the input words that were input, adding the output words to the input words until prescribed conditions are reached and outputting from document data words related to the input words;

a gathered-by-category word memory means for storing words gathered by the dictionary growth process by category;

a boundary word identification means for identifying boundary words belonging to multiple categories out of the words gathered by the dictionary growth process;

a category membership degree calculation means for calculating a category membership degree indicating the extent to which a boundary word belongs to the categories for each category to which the boundary word belongs, so that the category membership degree may become high, when the boundary word turns into an input word of the category, or when the boundary word turns into an output word of the category, on the basis of the information recorded in the input/output process recording means; and

a category update means for determining categories to which the boundary words belong on the basis of category membership degrees calculated by the category membership degree calculation means, and updating information stored in the gathered-by-category word memory means so as to reflect the determination results.

2. The dictionary creation device of claim 1, further comprising:

a dictionary growth means for gathering words in each category by repeatedly receiving input of words in each category, outputting from document data words related to the input words that were input, adding the output words to the input words until prescribed conditions are reached and outputting from document data words related to the input words.

3. The dictionary creation device of claim 1, wherein the input/output process recording means records information indicating an input/output process of input words and output words output by said input words, that repeats input/output multiple times.

4. The dictionary creation device according to claim 1, wherein the category membership calculation means calculates category membership degrees of the boundary words with respect to the categories on the basis of the shortest arrival distance until the first input words in the dictionary growth process in categories to which the boundary words belong arrive at the boundary words by repeating input/output.

5. The dictionary creation device of claim 4, wherein the category membership degree calculation means calculates category membership degrees of the boundary words with respect to the categories on the basis of a category closeness indicating the extent to which words in categories to which the boundary words belong are linked.

6. The dictionary creation device of claim 5, wherein the category membership degree calculation means finds the category closeness on the basis of the average value of the shortest arrival distance until the first input words in the dictionary growth process in categories to which the boundary words belong arrive at each word in the category by repeating input/output.

7. The dictionary creation device of claim 5, wherein the category membership degree calculation means finds the category closeness on the basis of the ratio of words having an input/output relationship to the words in the category to which the boundary words belong.

8. The dictionary creation device of claim 1, further comprising:

a means for updating the gathered-by-category word memory means so that only words within a useful distance out of the words in the category become words within the category, the useful distance being the shortest arrival distance until the first input words in the dictionary growth process within the category arrive at the boundary words by repeating input/output, for categories to which the category update means has determined that the boundary words belong.

9. The dictionary creation device of claim 1, further comprising:

a means for updating the gathered-by-category word memory means so that words more than a non-useful distance out of the words in the category are removed from the words in said category, the non-useful distance being the shortest arrival distance until the first input words in the dictionary growth process within the category arrive at the boundary words by repeating input/output, for categories to which the category update means has determined that the boundary words do not belong.

10. The dictionary creation device of claim 1, further comprising a cluster classification means for classifying words gathered by the dictionary growth process into clusters on the basis of information recorded in the input/output process recording means;

wherein the category membership calculation means finds the category membership degree for each word in a boundary cluster containing the boundary words and sets the average value or minimum value thereof as a category membership degree for the boundary cluster; and

the category update means determines categories to which words in the boundary clusters belong on the basis of the category membership degree of the boundary cluster, and updates information stored in the gathered-by-category word memory means so as to reflect the determination results.

11. The dictionary creation device of claim 10, wherein the cluster classification means calculates a degree of affinity between words indicating a value that becomes larger among words that are input by common words or among words that output common words out of the words gathered by the dictionary growth process, from information recorded in the input/output process recording means, and classifies words into clusters on the basis of the calculated degree of affinity.

12. A word gathering method comprising:

an input/output process recording step for recording information indicating an input/output process for input words and output words output by said input words, in a dictionary growth process for gathering words in each category by repeatedly receiving input of words in each category, outputting from document data words related to the input words that were input, adding the output words to the input words until prescribed conditions are reached and outputting from document data words related to the input words;

a gathered-by-category word memory step for storing words gathered by the dictionary growth process by category;

a boundary word identification step for identifying boundary words belonging to multiple categories out of the words gathered by the dictionary growth process;

a category membership degree calculation step for calculating a category membership degree indicating the extent to which the boundary words belong to the categories for each category to which the boundary words belong, so that the category membership degree may become high, when the boundary word turns into an input word of the category, or when the boundary word turns into an output word of the category, on the basis of the information recorded in the input/output process recording step; and

a category update step for determining categories to which the boundary words belong on the basis of category membership degrees calculated by the category membership degree calculation step, and updating information stored in the gathered-by-category word memory step so as to reflect the determination results.

13. A computer-readable recording medium on which is recorded a program that causes a computer to function as:

an input/output process recording means for recording information indicating an input/output process for input words and output words output by said input words, in a dictionary growth process for gathering words in each category by repeatedly receiving input of words in each category, outputting from document data words related to the input words that were input, adding the output words to the input words until prescribed conditions are reached and outputting from document data words related to the input words;

a gathered-by-category word memory means for storing words gathered by the dictionary growth process by category;

a boundary word identification means for identifying boundary words belonging to multiple categories out of the words gathered by the dictionary growth process;

a category membership degree calculation means for calculating a category membership degree indicating the extent to which the boundary words belong to the categories for each category to which the boundary words belong, so that the category membership degree may become high, when the boundary word turns into an input word of the category, or when the boundary word turns into an output word of the category, on the basis of the information recorded in the input/output process recording means; and

a category update means for determining categories to which the boundary words belong on the basis of category membership degrees calculated by the category membership degree calculation means, and updating information stored in the gathered-by-category word memory means so as to reflect the determination results.