SENTIMENT PREDICTION FROM TEXTUAL DATA

• US 20110112825A1
• Filed: 11/12/2009
• Published: 05/12/2011
• Est. Priority Date: 11/12/2009
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Bellegarda, Jerome
• Granted Patent: US 8,682,649 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G10L 13/10 Prosody rules derived from text; Stress or intonation

Abstract

A semantically organized domain space is created from a training corpus. Affective data are mapped onto the domain space to generate affective anchors for the domain space. A sentiment associated with an input text is determined based the affective anchors. A speech output may be generated from the input text based on the determined sentiment.

Claims

1. A method to train data for sentiment prediction, comprising:

creating a domain space that is a semantic representation of one or more identified areas of information; and

mapping affective data onto the domain space to generate first affective anchors for the domain space.

2. The method of claim 1, wherein the domain space is a vector space.

3. The method of claim 1 wherein the mapping the affective data to the domain space comprises

determining a first set of words associated with emotions;

forming word clusters within the domain space based on the first set of words; and

generating the first affective anchors based on the word clusters, wherein at least one of the first affective anchors is aligned with an emotion in the domain space.

4. The method of claim 1 wherein the mapping the affective data to the domain space comprises

creating an affective space from affective data;

determining second affective anchors for the affective space;

mapping at least a portion of the affective space onto the domain space;

wherein the first affective anchors for the domain space are generated based on locations in the domain space to which the second affective anchors are mapped.

5. The method of claim 4, wherein the affective space is a vector space.

6. The method of claim 1, wherein the domain space is created using latent semantic mapping (LSM).

7. The method of claim 4, wherein the affective space has a lower dimension than the domain space.

8. A computer readable storage medium having instructions stored thereon that, when executed, cause a computer to

create a domain space that is a semantic representation of one or more identified areas of information;

and

to map affective data onto the domain space to generate first affective anchors for the domain space.

9. The computer-readable storage medium of claim 8, wherein the domain space is a vector space.

10. The computer readable storage medium of claim 8, wherein the mapping the affective data to the domain space comprises

determining a first set of words associated with emotions;

forming word clusters within the domain space based on the first set of words; and

generating the first affective anchors based on the word clusters, wherein at least one of the first affective anchors is aligned with an emotion in the domain space.

11. The computer readable storage medium of claim 8 wherein the mapping the affective data to the domain space comprises

creating an affective space from affective data;

determining second affective anchors for the affective space;

mapping at least a portion of the affective space onto the domain space wherein the first

affective anchors for the domain space are generated based on locations in the domain space to which the second affective anchors are mapped.

12. The computer readable storage medium of claim 11, wherein the affective space is a vector space.

13. The computer readable storage medium of claim 8, wherein the domain space is created using latent semantic mapping (LSM).

14. The computer readable storage medium of claim 11, wherein the affective space has a lower dimension than the domain space.

15. A system for sentiment prediction, comprising:

a memory to store data associated with one or more identified areas of information;

a processor coupled to the memory;

a latent semantic mapping (LSM) module to create a domain space that is a semantic representation of one or more identified areas of information; and

a mapping module to map affective data onto the domain space to generate affective anchors for the domain space.

16. The system of claim 15, wherein the LSM module and the mapping module are stored in the memory wherein the mapping module further comprises:

a folding module to

determine a first set of words associated with emotions;

form word clusters within the domain space based on the first set of words; and

generate the first affective anchors based on the word clusters, wherein at least one of the first affective anchors is aligned with an emotion in the domain space.

17. The system of claim 15, the LSM module further to create an affective space from affective data and wherein the mapping module further comprises:

an embedding module to

determine second affective anchors for the affective space; and

to

map at least a portion of the affective space onto the domain space, wherein the first affective anchors for the domain space are generated based on locations in the domain space to which the second affective anchors are mapped.

18. The system of claim 15, wherein the domain space is a first vector space.

19. The system of claim 17, wherein the affective space has a lower dimension than the domain space.

20. A method to predict sentiment, comprising:

receiving a first text;

determining a representation of the first text in a domain space that provides a context associated with one or more second texts, wherein the domain space has one or more affective anchors representing one or more emotional categories; and

determining a sentiment associated with the first text based on the affective anchors.

21. The method of claim 20, further comprising

determining a closeness of the representation of the first text to the one or more affective anchors in the domain space.

22. The method of claim 20, further comprising

generating a speech output from the first text based on the sentiment.

23. A machine readable storage medium having instructions stored thereon that, when executed, cause a data processing system to perform operations comprising:

receiving a first text;

determining a representation of the first text in a domain space that provides a context associated with one or more second texts, wherein the domain space has one or more affective anchors representing emotional categories; and

determining a sentiment associated with the first text based on the one or more affective anchors.

24. The machine-readable storage medium of claim 23, further comprising instructions that cause the computer to perform operations comprising

determining a closeness of the representation of the first text to the one or more affective anchors in the domain space.

25. The computer-readable storage medium of claim 23, further comprising instructions that cause the computer to perform operations comprising

generating a speech output from the first text based on the sentiment.

26. A system to predict sentiment, comprising:

a processor;

an input to receive a first text;

a memory coupled to the processor;

a latent semantic mapping (LSM) module, stored in the memory, to determine a representation of the first text in a domain space that provides a context associated with one or more second texts, wherein the domain space has one or more affective anchors representing emotional categories; and

a sentiment computation module to determine a sentiment associated with the first text based on the one or more affective anchors.

27. The system of claim 26, wherein the LSM module is further to determine a closeness of the representation of the first text to the one or more affective anchors in the domain space.

28. The system of claim 27, further comprising

an output to generate a speech from the first text based on the sentiment, the output coupled to the processor.