Call transfer support system

Patent metadata

Field	Value
Publication number	US-11057526-B2
Application number	US-201916691172-A
Country	US
Kind code	B2
Filing date	Nov 21, 2019
Priority date	Nov 21, 2019
Publication date	Jul 6, 2021
Grant date	Jul 6, 2021

Abstract

A computer retrieves a dialog information records of the active call of the first operator. The computer extracts features from the dialog information records. The computer determines a feature vector from the extracted features and determines a transfer probability value based on the feature vector and previous call transfers to the second operator.

First claim

What is claimed is:

1. A processor-implemented method for transferring an active call from a first operator to a second operator, the method comprising:
retrieving dialog information records of the active call of the first operator;
extracting features from the dialog information records;
determining a feature vector from the extracted features, wherein the feature vector comprises values corresponding to one or more text features and one or more temporal features;
determining a transfer probability value based on the feature vector and previous call transfers to the second operator, wherein the transfer probability value is determined by summarizing each dimension of the feature vector multiplied by a corresponding weight to each dimension, and wherein the corresponding weight is determined by a transfer prediction model; and transferring the active call to the second operator based on determining the transfer probability value is above a threshold value.

2. The method of claim 1 , wherein determining the transfer probability value based on the feature vector and previous call transfers to the second operator comprises:
extracting features from the dialog information records of each previous call to the second operator;
determining a feature vector from the extracted features;
updating the transfer prediction model, wherein the transfer prediction model comprises a set of weights updated for each value of the feature vector; and
determining the transfer probability value based on applying the feature vector of the active call to the transfer prediction model.

3. The method of claim 2 , wherein updating the transfer prediction model is by a logistic regression when the feature vector has elements that are independently effective and by a deep neural network when the elements have combinatorial relationships.

4. The method of claim 1 , wherein extracting features from the dialog information records is by deep neural network.

5. The method of claim 1 , wherein the dialog information records are converted using text-to-speech dialogs with a corresponding timestamp.

6. The method of claim 5 , wherein the feature vector comprises text features and temporal features, wherein the text features are extracted from the dialog information records and temporal features are extracted from a time log.

7. A computer system for transferring an active call from a first operator to a second operator, the computer system comprising:
one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more memories, wherein the computer system is capable of performing a method comprising:
retrieving dialog information records of the active call of the first operator;
extracting features from the dialog information records;
determining a feature vector from the extracted features, wherein the feature vector comprises values corresponding to one or more text features and one or more temporal features;
determining a transfer probability value based on the feature vector and previous call transfers to the second operator, wherein the transfer probability value is determined by summarizing each dimension of the feature vector multiplied by a corresponding weight to each dimension, and wherein the corresponding weight is determined by a transfer prediction model; and transferring the active call to the second operator based on determining the transfer probability value is above a threshold value.

8. The computer system of claim 7 , wherein determining the transfer probability value based on the feature vector and previous call transfers to the second operator comprises:
extracting features from the dialog information records of each previous call to the second operator;
determining a feature vector from the extracted features;
updating the transfer prediction model, wherein the transfer prediction model comprises a set of weights updated for each value of the feature vector; and
determining the transfer probability value based on applying the feature vector of the active call to the transfer prediction model.

9. The computer system of claim 8 , wherein updating the transfer prediction model is by a logistic regression when the feature vector has elements that are independently effective and by a deep neural network when the elements have combinatorial relationships.

10. The computer system of claim 7 , wherein extracting features from the dialog information records is by deep neural network.

11. The computer system of claim 7 , wherein the dialog information records are converted using text-to-speech dialogs with a corresponding timestamp.

12. The computer system of claim 11 , wherein the feature vector comprises text features and temporal features, wherein the text features are extracted from the dialog information records and temporal features are extracted from a time log.

13. A computer program product for transferring an active call from a first operator to a second operator, the computer program product comprising:
one or more computer-readable tangible storage medium and program instructions stored on at least one of the one or more tangible storage medium, the program instructions executable by a processor, the program instructions comprising:
program instructions to retrieve dialog information records of the active call of the first operator;
program instructions to extract features from the dialog information records;
program instructions to determine a feature vector from the extracted features, wherein the feature vector comprises values corresponding to one or more text features and one or more temporal features;
program instructions to determine a transfer probability value based on the feature vector and previous call transfers to the second operator, wherein the transfer probability value is determined by summarizing each dimension of the feature vector multiplied by a corresponding weight to each dimension, and wherein the corresponding weight is determined by a transfer prediction model; and program instructions to transfer the active call to the second operator based on determining the transfer probability value is above a threshold value.

14. The computer program product of claim 13 , wherein program instructions to determine the transfer probability value based on the feature vector and previous call transfers to the second operator comprises:
program instructions to extract features from the dialog information records of each previous call to the second operator;
program instructions to determine a feature vector from the extracted features;
program instructions to update the transfer prediction model, wherein the transfer prediction model comprises a set of weights updated for each value of the feature vector; and
program instructions to determine the transfer probability value based on applying the feature vector of the active call to the transfer prediction model.

15. The computer program product of claim 14 , wherein program instructions to update the transfer prediction model is by a logistic regression when the feature vector has elements that are independently effective and by a deep neural network when the elements have combinatorial relationships.

16. The computer program product of claim 13 , wherein program instructions to extract features from the dialog information records is by deep neural network.

17. The compute

Assignees

• IBM

Inventors

• Takeuchi Emiko
• Kabeya Yoshinori
• Takuma Daisuke

Classifications

• G06F40/35 Primary

  Discourse or dialogue representation · CPC title

• H04M3/5237

  Interconnection arrangements between ACD systems · CPC title

• H04M3/58 Primary

  Arrangements for transferring received calls from one subscriber to another; Arrangements affording interim conversations between either the calling or the called party and a third party (substation line holding circuits H04M1/80) · CPC title

• H04M3/5175

  Call or contact centers supervision arrangements · CPC title

• H04M3/5233

  Operator skill based call distribution · CPC title