Domain-specific sentiment classification

• US 8,356,030 B2
• Filed: 06/17/2011
• Issued: 01/15/2013
• Est. Priority Date: 08/23/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Neylon, Tyler J., Wells, Michael, Reynar, Jeffrey C., Hannan, Kerry L., McDonald, Ryan T.
• Primary Examiner(s): Spieler, William
• Application Number:
  US13/163,623
• Publication Number:
  US 20110252036A1
• Time in Patent Office: 578 Days
• Field of Search: 707/721
• US Class Current: 707/721
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing

Abstract

A domain-specific sentiment classifier that can be used to score the polarity and magnitude of sentiment expressed by domain-specific documents is created. A domain-independent sentiment lexicon is established and a classifier uses the lexicon to score sentiment of domain-specific documents. Sets of high-sentiment documents having positive and negative polarities are identified. The n-grams within the high-sentiment documents are filtered to remove extremely common n-grams. The filtered n-grams are saved as a domain-specific sentiment lexicon and are used as features in a model. The model is trained using a set of training documents which may be manually or automatically labeled as to their overall sentiment to produce sentiment scores for the n-grams in the domain-specific sentiment lexicon. This lexicon is used by the domain-specific sentiment classifier.

Claims

1. A method of constructing a domain-specific sentiment classifier for classifying sentiment expressed by documents in a specified domain, comprising:

scoring sentiments expressed by one or more domain-specific documents, the scoring comprising;

determining that one or more of the domain-specific documents include an n-gram,

calculating a score for the n-gram included in the one or more domain-specific documents, and

calculating a sentiment score for the one or more domain-specific documents based on the score for the n-gram included in the documents;

creating a domain-specific sentiment lexicon based at least in part on said scoring sentiments expressed by one or more domain-specific documents;

generating the domain-specific sentiment classifier based on the domain-specific sentiment lexicon; and

storing the domain-specific sentiment classifier.

2. The method of claim 1, wherein scoring sentiments expressed by one or more domain-specific documents is based on a domain-independent sentiment lexicon, the method further comprising:

establishing the domain-independent sentiment lexicon by specifying a magnitude and polarity of sentiment expressed by each of a plurality of n-grams drawn from a domain-independent source.

3. The method of claim 1, wherein scoring sentiments expressed by one or more domain-specific documents is based on a domain-independent sentiment lexicon, the domain-independent sentiment lexicon specifies a magnitude and polarity of sentiment expressed by each of a plurality of n-grams, and wherein scoring sentiment expressed by one or more of the domain-specific documents based on a domain-independent sentiment lexicon comprises:

determining whether the n-gram included in the one or more documents is in the domain-independent sentiment lexicon.

4. The method of claim 3, wherein the score calculated for the n-gram included in the one or more documents is calculated based on one or more factors/techniques from the set consisting of:

a score for the n-gram specified in the domain-independent sentiment lexicon;

part-of-speech tagging based on a part-of-speech represented by the n-gram within the documents;

detecting whether the n-gram is used in a negative manner within the documents;

a location of the n-gram in the documents; and

stemming to identify a root of an n-gram in the documents.

5. The method of claim 1, further comprising:

identifying a set of high-sentiment documents based on scoring sentiments expressed by one or more domain-specific documents;

wherein creating a domain-specific sentiment lexicon is based at least in part on the set of high sentiment documents.

6. The method of claim 5, wherein identifying the set of high-sentiment documents based on scoring sentiments expressed by one or more domain-specific documents comprises:

identifying a set of high-positive-sentiment documents; and

identifying a set of high-negative-sentiment documents.

7. The method of claim 1, further comprises:

identifying a set of high-sentiment documents based on scoring sentiments expressed by one or more domain-specific documents;

wherein creating the domain-specific sentiment lexicon comprises;

identifying n-grams occurring within the documents in the set of high-sentiment documents;

filtering the identified n-grams to remove n-grams unlikely to express sentiment in the specific specified domain; and

storing the filtered n-grams as the domain-specific sentiment lexicon.

8. The method of claim 1, wherein generating the domain-specific sentiment classifier based on the domain-specific sentiment lexicon comprises:

building a model having n-grams of the domain-specific sentiment lexicon as features; and

training the model on a training corpus having a set of high-sentiment documents.

9. The method of claim 1, wherein generating the domain-specific sentiment classifier based on the domain-specific sentiment lexicon comprises:

building a model having n-grams of the domain-specific sentiment lexicon as features; and

training the model on a training corpus having domain-specific documents having manually-labeled sentiment scores.

10. The method of claim 9, wherein the model comprises a maximum entropy model.

11. The method of claim 9, wherein the training generates sentiment scores for the n-grams of the domain-specific sentiment lexicon and further comprising:

storing the sentiment scores for the n-grams of the domain-specific sentiment lexicon with the domain-specific sentiment lexicon.

12. The method of claim 1, further comprising:

using the domain-specific sentiment classifier to score a magnitude and polarity of sentiment expressed by a document in the specified domain.

13. A system for constructing a domain-specific sentiment classifier for classifying sentiment expressed by documents in a specified domain, comprising:

a non-transitory computer readable storage medium storing computer-executable instructions, the computer executable instructions comprising;

a scoring module configured to score sentiments expressed by one or more domain-specific documents, the scoring module configured to score by performing steps comprising;

determining that one or more of the domain-specific documents include an n-gram,

calculating a score for the n-gram included in the one or more domain-specific documents, and

calculating a sentiment score for the one or more domain-specific documents based on the score for the n-gram included in the documents;

a lexicon module configured to create a domain-specific sentiment lexicon based at least in part on said scoring sentiments expressed by one or more domain-specific documents;

a classifier module configured to generate the domain-specific sentiment classifier based on the domain-specific sentiment lexicon; and

a storing module configured to store the domain-specific sentiment classifier.

14. The system of claim 13, wherein scoring sentiments expressed by one or more domain-specific documents is based on a domain-independent sentiment lexicon, the domain-independent sentiment lexicon specifies a magnitude and polarity of sentiment expressed by each of a plurality of n-grams, and the scoring module is further configured to:

determine whether the n-gram included in the one or more documents is in the domain-independent sentiment lexicon.

15. The system of claim 13, further comprising:

an analysis module configured to identify a set of high-sentiment documents based on scoring sentiments expressed by one or more domain-specific documents;

wherein the lexicon module is configured to create the domain-specific sentiment lexicon based at least in part on the set of high-sentiment documents.

16. The system of claim 15, wherein the lexicon module is further configured to:

identify n-grams occurring within the documents in the set of high-sentiment documents;

filter the identified n-grams to remove n-grams unlikely to express sentiment in the specified domain; and

store the filtered n-grams as the domain-specific sentiment lexicon.

17. The system of claim 13, wherein the classifier module is further configured to:

build a model having n-grams of the domain-specific sentiment lexicon as features; and

train the model on a training corpus having a set of high-sentiment documents.

18. The system of claim 13, wherein the classifier module is further configured to:

build a model having n-grams of the domain-specific sentiment lexicon as features; and

train the model on a training corpus having domain-specific documents having manually-labeled sentiment scores.

19. The system of claim 18, wherein the model comprises a maximum entropy model.

20. The system of claim 18, wherein the training generates sentiment scores for the n-grams of the domain-specific sentiment lexicon and wherein the storing module is further configured to:

store the sentiment scores for the n-grams of the domain-specific sentiment lexicon with the domain-specific sentiment lexicon.

21. The system of claim 13, wherein the scoring module is further configured to:

score a magnitude and polarity of sentiment expressed by a document in the specified domain using the domain-specific sentiment classifier.

22. A computer program product having a computer-readable storage medium having computer-executable code for constructing a domain-specific sentiment classifier for classifying sentiment expressed by documents in a specified domain, the computer-executable code comprising:

a scoring module configured to score sentiments expressed by a plurality of domain-specific documents, the scoring module configured to score by performing steps comprising;

determining that one or more of the domain-specific documents include an n-gram,

calculating a score for the n-gram included in the one or more domain-specific documents, and

calculating a sentiment score for the one or more domain-specific documents based on the score for the n-gram included in the documents;

a lexicon module configured to create a domain-specific sentiment lexicon based at least in part on said scoring sentiments expressed by one or more domain-specific documents;

a classifier module configured to generate the domain-specific sentiment classifier based on the domain-specific sentiment lexicon; and

a storing module configured to store the domain-specific sentiment classifier.

23. The computer program product of claim 22, wherein scoring sentiments expressed by one or more domain-specific documents is based on a domain-independent sentiment lexicon, the domain-independent sentiment lexicon specifies a magnitude and polarity of sentiment expressed by each of a plurality of n-grams, and the scoring module is further configured to:

determine whether the n-gram included in the one or more documents is in the domain-independent sentiment lexicon.

24. The computer program product of claim 22, further comprising:

an analysis module configured to identify a set of high-sentiment documents based on scoring sentiments expressed by one or more domain-specific documents;

wherein the lexicon module is configured to create the domain-specific sentiment lexicon based at least in part on the set of high-sentiment documents.

25. The computer program product of claim 24, wherein the lexicon module is further configured to:

identify n-grams occurring within the documents in the set of high-sentiment documents;

filter the identified n-grams to remove n-grams unlikely to express sentiment in the specified domain; and

store the filtered n-grams as the domain-specific sentiment lexicon.

26. The computer program product of claim 22, wherein the classifier module is further configured to:

build a model having n-grams of the domain-specific sentiment lexicon as features; and

train the model on a training corpus having a set of high-sentiment documents.

27. The computer program product of claim 22, wherein the classifier module is further configured to:

build a model having n-grams of the domain-specific sentiment lexicon as features; and

train the model on a training corpus having domain-specific documents having manually-labeled sentiment scores.

28. The computer program product of claim 27, wherein the model comprises a maximum entropy model.

29. The computer program product of claim 27, wherein the training generates sentiment scores for the n-grams of the domain-specific sentiment lexicon and wherein the storing module is further configured to:

store the sentiment scores for the n-grams of the domain-specific sentiment lexicon with the domain-specific sentiment lexicon.

30. The computer program product of claim 22, wherein the scoring module is further configured to:

score a magnitude and polarity of sentiment expressed by a document in the specified domain using the domain-specific sentiment classifier.