System and Methods for Evaluating Feature Opinions for Products, Services, and Entities

• US 20090265332A1
• Filed: 04/18/2008
• Published: 10/22/2009
• Est. Priority Date: 04/18/2008
• Status: Active Grant

• Current Assignee: inContact Incorporated (Nice Ltd)
• Original Assignee: Biz360, Inc. (Attensity Group Incorporated)
• Inventors: Mushtaq, Faisal, Das, Gavin, Habib, Adnan Mohsin
• Granted Patent: US 8,117,207 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06Q 30/0203 Market surveys; Market polls
• CPC Class Codes: G06Q 30/0245 Surveys

Abstract

A system for evaluating a review having unstructured text comprises a segment splitter for separating at least a portion of the unstructured text into one or more segments, each segment comprising one or more words; a segment parser coupled to the segment splitter for assigning one or more lexical categories to one or more of the one or more words of each segment; an information extractor coupled to the segment parser for identifying a feature word and an opinion word contained in the one or more segments; and a sentiment rating engine coupled to the information extractor for calculating an opinion score based upon an opinion grouping, the opinion grouping including at least the feature word and the opinion word identified by the information extractor.

Claims

1. A system for evaluating a review having unstructured text, the system comprising:

a segment splitter for separating at least a portion of the unstructured text into one or more segments, each segment comprising one or more words;

a segment parser coupled to the segment splitter for assigning one or more lexical categories to one or more of the one or more words of each segment;

an information extractor coupled to the segment parser for identifying a feature word and an opinion word contained in the one or more segments; and

a sentiment rating engine coupled to the information extractor for calculating an opinion score based upon an opinion grouping, the opinion grouping including at least the feature word and the opinion word identified by the information extractor.

2. The system of claim 1, further comprising a language engine coupled to the segment parser for providing the segment parser with data linking the one or more words of each segment to the lexical categories to which the one or more words are customarily assigned in a dictionary.

3. The system of claim 1, wherein the information extractor is configured to identify rating modifiers contained in the one or more segments.

4. The system of claim 1, further comprising a training model coupled to the information extractor for providing the information extractor with a set of sample segments and corresponding trained structures to assist in identifying feature words and opinion words.

5. The system of claim 1, further comprising a nonsense analyzer coupled to the sentiment rating engine for determining whether the opinion grouping likely expresses a valid opinion.

6. The system of claim 5, wherein the nonsense analyzer is configured to determine whether the opinion grouping duplicates an earlier opinion grouping.

7. The system of claim 1, further comprising a modifier tool coupled to the sentiment rating engine for assigning word scores to the feature word and the opinion word contained in the opinion grouping.

8. The system of claim 1, further comprising an opinion insight summarizer coupled to the sentiment rating engine, wherein

the sentiment rating engine calculates a plurality of opinion scores based upon a plurality of opinion groupings;

and

the opinion insight summarizer aggregates the plurality of opinion scores into one or more advocacy scores.

9. The system of claim 8, further comprising a feature/attribute management tool coupled to the opinion insight summarizer for assigning a relative weight to each of the plurality of opinion scores.

10. The system of claim 8, further comprising a presentation engine coupled to the opinion insight summarizer for presenting the one or more advocacy scores to a user on an output device.

11. A method for evaluating a review having unstructured text, the method comprising the steps of:

separating at least a portion of the unstructured text into one or more segments, each segment comprising one or more words;

assigning lexical categories to one or more of the one or more words of each segment;

identifying a feature word and an opinion word contained in the one or more segments; and

calculating an opinion score based upon an opinion grouping, the opinion grouping including the feature word and the opinion word.

12. The method of claim 11, wherein the step of assigning lexical categories further comprises the step of providing data linking the one or more words of each segment to the lexical categories to which the one or more words are customarily assigned in a dictionary.

13. The method of claim 11, wherein the step of identifying feature words and opinion words further comprises the step of identifying rating modifiers contained in the one or more segments.

14. The method of claim 11, wherein the step of identifying feature words and opinion words further comprises the step of providing a set of sample segments and corresponding trained structures to assist in identifying feature words and opinion words.

15. The method of claim 11, wherein the step of calculating an opinion score further comprises the step of determining whether the opinion grouping likely expresses a valid opinion.

16. The method of claim 11, wherein the step of calculating an opinion score further comprises the step of determining whether the opinion grouping duplicates an earlier opinion grouping.

17. The method of claim 11, wherein the step of calculating an opinion score further comprises the step of assigning word scores to the feature word and the opinion word contained in the opinion grouping.

18. The method of claim 11, further comprising the steps of

calculating a plurality of opinion scores based upon a plurality of opinion groupings;

and

aggregating the plurality of opinion scores into one or more advocacy scores.

19. The method of claim 18, wherein the step of aggregating the plurality of opinion scores further comprises assigning a relative weight to each of the plurality of opinion scores.

20. The method of claim 18, further comprising the step of presenting the one or more advocacy scores to a user on an output device.

21. A system for evaluating a review having structured text, the system comprising:

a review mapper for mapping the review to one or more of a product, service, or entity;

a product mapper coupled to the review mapper for separating the structured text into one or more constituent parts; and

a normalizer tool coupled to the product mapper for normalizing the one or more constituent parts to a predetermined format.

22. The system of claim 21, further comprising an operator interface coupled to the review mapper for allowing a user to manually map the review to one or more of a product, service, or entity.

23. A method for creating a training model for a system for evaluating a review having unstructured text, the method comprising the steps of:

displaying a series of text segments on an output device, each text segment comprising one or more words;

receiving an identification of the words in each text segment that correspond to a feature or to an opinion; and

creating a set of sample text segments and corresponding trained structures based upon the identification.