Method of identifying topic of text using nouns

• US 7,805,291 B1
• Filed: 05/25/2005
• Issued: 09/28/2010
• Est. Priority Date: 05/25/2005
• Status: Expired due to Fees

• Current Assignee: National Security Agency
• Original Assignee: The United States of America As Represented By The Director National Security Agency
• Inventors: Berkowitz, Sidney
• Primary Examiner(s): Wozniak; James S
• Assistant Examiner(s): Adesanya; Olujimi A
• Application Number:
  US11/137,594
• Time in Patent Office: 1,952 Days
• Field of Search: 704/4, 704/7, 704/10
• US Class Current: 704/9
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning

Abstract

A method of identifying a topic of a text. Text is received. Then, the nouns in the text are identified. The singular form of each identified noun is determined. Combinations are created of the singular form of the identified nouns, where the number of singular forms of the nouns in the combinations is user-definable. The frequency of occurrence in the text of each noun that corresponds to its singular form is determined. Each frequency of occurrence is assigned as a score to its corresponding singular form noun. Each combination of singular form nouns is assigned a score that is equal to the sum of the scores of its constituent singular form nouns. The user-definable number of top scoring singular form nouns and combinations of singular form nouns are selected as the topic of the text.

Claims

1. A method of identifying a topic of a text, comprising the steps of

a) receiving the text, where the text includes words;

b) identifying in a hardware device each unique word in the text that is a noun;

c) determining in the hardware device a singular form of each noun identified in step (b);

d) creating in the hardware device combinations of the singular forms of the nouns determined in step (c), where the number of singular forms of the nouns in each combination is a user-definable;

e) determining in the hardware device a frequency of occurrence in the text of each noun identified in step (b);

f) assigning in the hardware device a score to each singular form noun, where the score of each singular form noun is the frequency of occurrence of the corresponding noun determined in step (e);

g) assigning in the hardware device a score to each combination of singular form nouns, where the score of each combination of singular form nouns is a sum of the scores of the singular form nouns in the combination;

h) selecting in the hardware device a user-definable number of scores of singular form nouns that are greatest in value;

i) selecting in the hardware device a user-definable number of scores of combinations of singular form nouns that are greatest in value; and

j) returning from the hardware device the singular forms of nouns and the combinations of singular forms nouns that correspond to the scores selected in step (h) and step (i) as the topic of the text.

2. The method of claim 1, wherein said step of receiving the text is comprised of the step of receiving the text, where the text is selected from the group of text consisting of speech-based text, optical-character-read text, and any other suitable and equivalent text.