Sentiment expression analysis based on keyword hierarchy

• US 9,477,704 B1
• Filed: 12/19/2013
• Issued: 10/25/2016
• Est. Priority Date: 12/31/2012
• Status: Active Grant

• Current Assignee: Teradata Corporation
• Original Assignee: Teradata US, Inc. (Teradata Corporation)
• Inventors: Kottu, Rajeswara R., Anderson, Joshua D., Lee, Harold R., Yapejian, Ara
• Primary Examiner(s): THAI, HANH B
• Application Number:
  US14/135,515
• Time in Patent Office: 1,041 Days
• Field of Search: 707/748, 707/752, 707/708, 707/48
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24 Querying

Abstract

A sentiment-scoring system may include a storage device configured to store a plurality of keywords, keyword groups, and a keyword group hierarchy. Each keyword may be associated with at least one of the keyword groups. The keyword hierarchy may include a hierarchy associated with each keyword group. The system may further include a processor in communication with the storage device. The processor may be configured to locate a plurality of sentiment expressions and identify a plurality of keywords present in the plurality of sentiment expressions. The processor may be further configured to determine at least one respective keyword group associated with each identified keyword and determine a sentiment score for each sentiment expression with respect to the associated keyword group within the keyword hierarchy. The processor may be further configured to provide at least one sentiment score to a display. A method and computer-readable medium may also be implemented.

Claims

1. A sentiment-scoring system, comprising:

a storage device configured to store a plurality of keywords, keyword groups, and a keyword group hierarchy, wherein each keyword is associated with at least one of the keyword groups, and wherein the keyword hierarchy comprises a hierarchy associated with each keyword group;

a processor in communication with the storage device, the processor configured to;

locate a plurality of sentiment expressions;

identify a plurality of keywords present in the plurality of sentiment expressions;

determine at least one respective keyword group associated with each identified keyword;

determine a sentiment score for each sentiment expression with respect to the associated at least one respective keyword group within the keyword hierarchy, wherein each sentiment score is indicative of a level of sentiment about the associated at least one respective keyword group; and

provide at least one determined sentiment score to a display.

2. The sentiment-scoring system of claim 1, wherein each keyword group includes a plurality of keyword subgroups, wherein each keyword subgroup is associated with at least one keyword, wherein at least one subgroup is configured to serve as a parent keyword subgroup to a plurality of other keyword subgroups, wherein each of the plurality of other subgroups may be associated with a respective keyword and respective keyword sentiment score indicative of a respective level of sentiment associated with the respective keyword, and wherein each of the respective keyword sentiment scores may be combined to create an overall sentiment score for the parent keyword subgroup.

3. The sentiment-scoring system of claim 2, wherein each of the other keyword subgroups is associated with a respective weighting factor, wherein each weighting factor is indicative of a relative level of impact of the associated sentiment.

4. The sentiment-scoring system of claim 3, wherein each of the respective weighting factors is a percentage with respect to the parent keyword subgroup.

5. The sentiment-scoring system of claim 3, wherein at least one of the respective weighting factors and the percentages are user-adjustable.

6. The sentiment-scoring system of claim 1, wherein the processor is further configured to locate a plurality of sentiment expressions on at least one social network.

7. The sentiment-scoring system of claim 1, wherein each of the other keyword subgroups may serve as a parent keyword subgroup to additional other keyword subgroups.

8. A method comprising:

locating, with a processor, a plurality of sentiment expressions;

identifying, with the processor, a plurality of predetermined keywords present in the plurality of sentiment expressions, wherein each keyword is included in at least one of a plurality of keyword groups, wherein each keyword group includes an associated keyword hierarchy;

determining, with the processor, at least one respective keyword group associated with each identified keyword;

determining, with the processor, a sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy, wherein each sentiment score is indicative of a level of sentiment about the associated at least one keyword group; and

providing, with the processor, at least one determined sentiment score to a display.

9. The method of claim 8, wherein determining, with the processor, at least one respective keyword group associated with each identified keyword comprises identifying, with the processor, a plurality of keyword subgroups included in at least one of the keyword group, wherein each keyword subgroup is associated with at least one keyword, wherein at least one subgroup is configured to serve as a parent keyword subgroup to a plurality of other subgroups, wherein each of the plurality of other subgroups may be associated with a respective keyword, wherein determining, with the processor, a sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy comprises determining, with the processor, a respective keyword sentiment score for each of the other keyword subgroups, wherein each of the respective keyword sentiment scores is indicative of a respective level of sentiment associated with the respective keyword, and wherein each of the respective keyword sentiment scores are combined to create an overall sentiment score for the parent keyword subgroup.

10. The method of claim 9, wherein each of the other keyword subgroups is associated with a respective weighting factor, wherein each weighting factor is indicative of a relative level of impact of the associated sentiment, and wherein determining, with the processor, a sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy comprises determining, with the processor, a respective keyword sentiment score for each of the other keyword subgroups based on the respective weight factor.

11. The method of claim 10, wherein each of the respective weighting factors is a percentage with respect to the parent keyword subgroup.

12. The method of claim 10, further comprising receiving, with the processor, user-input to adjust at least one of the respective weighting factors.

13. The method of claim 8, wherein locating, with a processor, the plurality of sentiment expressions comprises locating, with a processor, a plurality of sentiment expressions on at least one social network.

14. The method of claim 8, wherein each of the other keyword subgroups may serve as a parent keyword subgroup to additional other keyword subgroups.

15. A computer-readable medium encoded with a plurality of instructions executable by a processor, the plurality of instructions comprising:

instructions to locate a plurality of sentiment expressions;

instructions to identify a plurality of predetermined keywords present in the plurality of sentiment expressions, wherein each keyword is included in at least one of a plurality of keyword groups, wherein each keyword group includes an associated keyword hierarchy;

instructions to determine at least one respective keyword group associated with each identified keyword;

instructions to determine a sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy, wherein each sentiment score is indicative of a level of sentiment about the associated at least one keyword group; and

instructions to provide at least one determined sentiment score to a display.

16. The computer-readable medium of claim 15, wherein the instructions to determine at least one respective keyword group associated with each identified keyword comprise instructions to identifying a plurality of keyword subgroups included in at least one of the keyword group, wherein each keyword subgroup is associated with at least one keyword, wherein at least one subgroup is configured to serve as a parent keyword subgroup to a plurality of other subgroups, wherein each of the plurality of other subgroups may be associated with a respective keyword, wherein the instructions to determine a sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy comprise instructions to determine a respective keyword sentiment score for each of the other keyword subgroups, wherein each of the respective keyword sentiment scores is indicative of a respective level of sentiment associated with the respective keyword, and wherein each of the respective keyword sentiment scores are combined to create an overall sentiment score for the parent keyword subgroup.

17. The computer-readable medium of claim 16, wherein each of the other keyword subgroups is associated with a respective weighting factor, wherein each weighting factor is indicative of a relative level of impact of the associated sentiment, and wherein the instructions to determine the sentiment score for each sentiment expression with respect to the associated at least one keyword group within the keyword hierarchy comprise instructions to determine a respective keyword sentiment score for each of the other keyword subgroups based on the respective weight factor.

18. The computer-readable medium of claim 17, wherein each of the respective weighting factors is a percentage with respect to the parent keyword subgroup.

19. The computer-readable medium of claim 17, wherein the plurality of instructions further comprise instructions to receive user-input to adjust at least one of the respective weighting factors.

20. The computer-readable medium of claim 15, wherein the instructions to locate the plurality of sentiment expressions comprise instructions to locate a plurality of sentiment expressions on at least one social network.