System and method for extracting demographic information

• US 20050169441A1
• Filed: 01/30/2004
• Published: 08/04/2005
• Est. Priority Date: 01/30/2004
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Simsk, Steven J., Yacoub, Sherif, Lin, Xiaofan, Vincent, Francois
• Granted Patent: US 7,349,527 B2
• Time in Patent Office: Days
• US Class Current: 379/88.160
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04M 2201/41 using speaker recognition
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/5166 in combination with interactive voice response systems or voice portals, e.g. as front-ends

Abstract

A system and method for extracting demographic information from a contact is disclosed. The method discloses: initiating a dialog between a contact and a call handling system; selecting a set of demographic characteristics; assigning a set of acoustic confidence scores; assigning a set of substantive confidence scores; combining the acoustic and substantive confidence scores; and tailoring information presented to the contact using the set of combined confidence scores. The system discloses: an Interactive Voice Response module for initiating a dialog between a contact and a call handling system, and selecting a set of demographic characteristics; an acoustic classifier for assigning a set of acoustic confidence scores; a substantive classifier for assigning a set of substantive confidence scores; and a data combiner for combining the acoustic and substantive confidence scores. The Interactive Voice Response module also tailors information presented to the contact using the set of combined confidence scores.

Claims

1. A method for extracting demographic information, comprising:

initiating a dialog between a contact and a call handling system;

selecting a set of demographic characteristics;

assigning a set of acoustic confidence scores to the demographic characteristics;

assigning a set of substantive confidence scores to the demographic characteristics;

combining the acoustic and substantive confidence scores for each of the demographic characteristics; and

tailoring information presented to the contact using the set of combined confidence scores.

2. The method of claim 1, wherein assigning substantive confidence scores includes:

presenting the contact with a first substantive dialog;

collecting a set of responses to the first substantive dialog from the contact;

comparing the contact'"'"'s responses to a predefined body of responses associated with the set of demographic characteristics; and

assigning a first set of substantive confidence scores to the demographic characteristics.

3. The method of claim 2, wherein presenting includes:

continuing to present the contact with the substantive dialog until one of the substantive dialog confidence score reaches a predetermined value.

4. The method of claim 2, wherein presenting includes:

continuing to present the contact with the substantive dialog until a predetermined time period has expired.

5. The method of claim 2, wherein presenting includes:

presenting the substantive dialog to the contact when the contact is placed on hold.

6. The method of claim 2, wherein assigning substantive confidence scores includes:

presenting the contact with a second substantive dialog, in response to a request from the call handling system;

collecting a set of responses to the second substantive dialog from the contact;

comparing the contact'"'"'s responses to the predefined body of responses associated with the set of demographic characteristics; and

assigning a second set of substantive confidence scores to the demographic characteristics.

7. The method of claim 1, wherein assigning substantive confidence scores includes:

presenting the contact with a probing dialog;

collecting a set of responses to the probing dialog from the contact;

comparing the contact'"'"'s responses to a predefined body of probing dialog responses associated with the set of demographic characteristics; and

assigning a set of probing dialog confidence scores to the demographic characteristics.

8. The method of claim 7, wherein presenting includes:

asking the contact a set of questions associated with the demographic characteristics.

9. The method of claim 1, wherein assigning substantive confidence scores includes:

presenting the contact with a set of multiple choice questions;

collecting a set of responses to the multiple choice questions from the contact;

comparing the contact'"'"'s responses to a predefined body of multiple choice question responses associated with the set of demographic characteristics; and

assigning a set of multiple choice confidence scores to the demographic characteristics.

10. The method of claim 9, wherein presenting includes:

presenting the contact with multiple choice questions associated with the demographic characteristics.

11. The method of claim 1, wherein assigning acoustic confidence scores includes:

extracting an acoustic feature from the contact'"'"'s speech signal; and

comparing the feature to a predefined body of speech signal features associated with the set of demographic characteristics.

12. The method of claim 1, wherein combining includes:

weighting the confidence scores using ground truth data.

13. The method of claim 1, wherein weighting includes:

adjusting a first confidence scores weight for a given demographic characteristic if the first confidence score differs from a second confidence score for that given demographic characteristic by a predetermined amount.

14. The method of claim 1, wherein combining includes:

multiplying together the confidence scores for each demographic characteristic.

15. The method of claim 1, wherein combining includes:

combining the confidence scores for each demographic characteristic according to the following formula;

$S<br><br>\left(C_i\right)=\underset{j=1}{\overset{N}{\sum <br><br>}}<br><br>r_j<br><br>p_{\mathrm{ij}}$

(where N is a total number of classifiers, Ci is the i'"'"'th demographic characteristic, and Pij is a confidence score for Ci generated by Classifier j, and r_j is trained weights);

16. The method of claim 1, wherein combining includes:

combining the confidence scores from each classifier for each demographic characteristic according to the following formula;

$S<br><br>\left(C_i\right)=\underset{j=1}{\overset{N}{\prod <br><br>}}<br><br>p_{\mathrm{ij}}^{r_j}$

(where N is a total number of classifiers, Ci is the i'"'"'th demographic characteristic, and Pij is a confidence score for Ci generated by Classifier j, and r_j is trained weights);

17. The method of claim 1, wherein combining includes:

using a neural net to combine the confidence scores for each demographic characteristic.

18. The method of claim 17, wherein the neural net is a Multiple Layer Perception (MLP) network.

19. The method of claim 1, wherein tailoring includes:

identifying a sub-set of the demographic characteristics having combined confidence scores exceeding a predetermined set of thresholds; and

presenting the contact with information specifically directed to contacts having the sub-set of demographic characteristics.

20. The method of claim 19, wherein the predetermined threshold is equal to a highest combined confidence score.

21. The method of claim 1, wherein the demographic characteristics include gender, age, accent, and stress level.

22. A method for extracting demographic information, comprising:

initiating a dialog between a contact and a call handling system;

selecting a set of demographic characteristics;

assigning a set of acoustic confidence scores to the demographic characteristics;

assigning a set of substantive confidence scores to the demographic characteristics;

combining the acoustic and substantive confidence scores for each of the demographic characteristics;

tailoring information presented to the contact using the set of combined confidence scores;

presenting the contact with a probing dialog;

collecting a set of responses to the probing dialog from the contact;

comparing the contact'"'"'s responses to a predefined body of probing dialog responses associated with the set of demographic characteristics;

assigning a set of probing dialog confidence scores to the demographic characteristics;

presenting the contact with a set of multiple choice questions;

collecting a set of responses to the multiple choice questions from the contact;

comparing the contact'"'"'s responses to a predefined body of multiple choice question responses associated with the set of demographic characteristics; and

assigning a set of multiple choice confidence scores to the demographic characteristics.

23. A computer-usable medium embodying computer program code for commanding a computer to extract demographic information, comprising:

initiating a dialog between a contact and a call handling system;

selecting a set of demographic characteristics;

assigning a set of acoustic confidence scores to the demographic characteristics;

assigning a set of substantive confidence scores to the demographic characteristics;

combining the acoustic and substantive confidence scores for each of the demographic characteristics; and

tailoring information presented to the contact using the set of combined confidence scores.

24. A system for extracting demographic information, comprising a:

means for initiating a dialog between a contact and a call handling system;

means for selecting a set of demographic characteristics;

means for assigning a set of acoustic confidence scores to the demographic characteristics;

means for assigning a set of substantive confidence scores to the demographic characteristics;

means for combining the acoustic and substantive confidence scores for each of the demographic characteristics; and

means for tailoring information presented to the contact using the set of combined confidence scores.

25. A system for extracting demographic information, comprising:

an Interactive Voice Response module for initiating a dialog between a contact and a call handling system, and selecting a set of demographic characteristics;

an acoustic classifier for assigning a set of acoustic confidence scores to the demographic characteristics;

a substantive classifier for assigning a set of substantive confidence scores to the demographic characteristics; and

a data combiner for combining the acoustic and substantive confidence scores for each of the demographic characteristics; and

wherein the Interactive Voice Response module further tailors information presented to the contact using the set of combined confidence scores.

26. The system of claim 25, wherein the substantive classifier includes:

a probing dialog classifier for assigning a set of probing dialog confidence scores to the demographic characteristics; and

a multiple choice classifier for assigning a set of multiple choice confidence scores to the demographic characteristics.