Method and apparatus for analyzing affect and emotion in text

• US 6,622,140 B1
• Filed: 11/15/2000
• Issued: 09/16/2003
• Est. Priority Date: 11/15/2000
• Status: Expired due to Fees

• Current Assignee: Justsystems Corporation
• Original Assignee: Justsystems Corporation
• Inventors: Kantrowitz, Mark
• Primary Examiner(s): Alam, Shahid Al
• Assistant Examiner(s): HWANG, JOON H
• Application Number:
  US09/713,726
• Time in Patent Office: 1,035 Days
• Field of Search: 707/1-10, 707/100-104, 707/500, 707/531, 704/1-10, 704/231, 704/251-257, 715/500.1, 715/531, 715/532
• US Class Current: 1/1
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99937 Sorting

Abstract

Disclosed is a computer-assisted method for classifying a text document according to emotion and affect. A score is assigned to each affect term in the document. An affect score is computed for the document from the scores for each affect term. The document is then classified in accordance with the affect score. An apparatus for performing the computer-assisted method is also disclosed.

Claims

1. A computer-assisted method for classifying a text document according to emotion and affect, comprising the steps of:

assigning a score to each affect term in the document,

computing an affect score for the document from the scores for each affect term,

classifying the document in accordance with the affect score,

classifying affect terms as positive or negative,

identifying named entities in the document,

assigning the scores of the affect terms to the named entities, and for each named entity,

summing the scores of all positive affect terms,

summing the scores of all negative affect terms, and

computing an affect score by subtracting the sum of the scores of all negative affect terms from the sum of the scores of all positive affect terms.

2. The computer-assisted method according to claim 1, wherein the affect terms include one or more consecutive words.

3. The computer-assisted method according to claim 1, wherein the score assigned to each affect term is assigned by counting the number of times each affect term occurs within the document and assigning the number of times the term occurs as the score.

4. The computer-assisted method according to claim 1, wherein the score assigned to each affect term is assigned by:

assigning an intensity value for each term, and

counting the number of times each affect term occurs within the document,

wherein the score assigned to each affect term is the number of times the term occurs multiplied by the intensity value for the terms as the score.

5. The computer-assisted method according to claim 1, wherein if a negation term is used in conjunction with the affect term, the positive affect term is treated as a negative affect term, and vice versa.

6. The computer-assisted method according to claim 1, wherein each affect term is classified as positive or negative at each occurrence of the term based on the part of speech of the affect term at the occurrence.

7. The computer-assisted method according to claim 1, wherein each affect term is classified as positive or negative at each occurrence of the affect term based on the meaning of the affect term at the occurrence.

8. The computer-assisted method according to claim 1, wherein the affect score is the sum of the scores of all positive affect terms and the sum of the scores of all negative affect terms.

9. The computer-assisted method according to claim 1, wherein the score assigned to each affect term is assigned by additionally parsing sentences of the document such that each verb'"'"'s score is assigned to the verb'"'"'s agent and objects and each modifier'"'"'s score is assigned to the modifier'"'"'s objects.

10. The computer-assisted method according to claim 1, wherein the classification of the affect term as positive or negative is based on the direction from the affect term to the named entity.

11. The computer-assisted method according to claim 1, wherein the assignment of the affect term score to the named entity occurs when the affect term and the named entity are in the same sentence.

12. The computer-assisted method according to claim 1, wherein the scores from the affect terms are assigned to the named entities by, for each affect term, assigning the score from the affect term to the closest named entity.

13. The computer-assisted method according to claim 1, wherein the scores from the affect terms are assigned to the named entities by, for each named entity, assigning the score from the closest affect term to the named entity.

14. The computer-assisted method according to claim 1, wherein the assignment occurs when no other named entity is between the named entity and the closest affect term.

15. The computer-assisted method according to claim 1, further comprising the step of:

canonicalizing variants of the named entities into groups of synonymous variants to be treated as a single named entity.

16. The computer-assisted method according to claim 1, wherein the text document is a news article.

17. The computer-assisted method according to claim 16, wherein the news article is a financial news article.

18. The computer-assisted method according to claim 1, wherein the text document is a web page.

19. The computer-assisted method according to claim 1, wherein the text document is customer communications.

20. An apparatus to enable a method for classifying a text document according to emotion and affect, comprising:

means for assigning a score to each affect term in the document,

means for computing an affect score for the document from the scores for each affect term,

means for classifying the document in accordance with the affect score,

means for classifying the affect terms as positive or negative,

means for identifying named entities in the document,

means for assigning the scores of the affect terms to the named entities, and for each named entity,

means for summing the scores of all positive affect terms,

means for summing the scores of all negative affect terms, and

means for computing an affect score by subtracting the sum of the scores of all negative affect terms from the sum of the scores of all positive affect terms.