CN-122024740-A - Risk control method, apparatus, device, storage medium, and computer program product

CN122024740ACN 122024740 ACN122024740 ACN 122024740ACN-122024740-A

Abstract

The application discloses a risk control method, device, equipment, storage medium and computer program product based on voiceprint recognition, which are used for solving the problem that the system efficiency is low because the traditional voiceprint verification scheme is directly applied to a telecom customer service scene and is not optimized for the telecom customer service scene. The method comprises the steps of determining a work order risk level corresponding to a customer service work order to be processed according to a predefined business risk rule base, obtaining historical behavior data of a user corresponding to the customer service work order to be processed, determining a customer risk label according to the historical behavior data, preprocessing original call voice when judging that the customer service work order to be processed carries out voiceprint verification according to the work order risk level and the customer risk label to obtain call voice to be analyzed, carrying out voiceprint matching on the call voice to be analyzed according to a voiceprint matching model to obtain confidence coefficient corresponding to the call voice to be analyzed, and executing a risk coping process according to a risk control strategy matched with the confidence coefficient.

Inventors

LI ZHUOPENG
FU YURONG
FAN JIN
MENG XIN
ZHANG YUJIE
SHI PENGHAO
WEI KAIJIN
YANG JIAN
LU YANHUI
WANG FENG
YAN SHAOHUA
GUO QIANG
WANG XINYUAN
Xia Shige
YUE TENGFEI
CHEN YUJIE
LI WEIWEI
JIN YI

Assignees

中移在线服务有限公司
中国移动通信集团有限公司

Dates

Publication Date: 20260512
Application Date: 20260121

Claims (10)

1. A risk control method based on voiceprint recognition, comprising: Determining a work order risk level corresponding to the customer service work order to be processed according to a predefined business risk rule base; Acquiring historical behavior data of a user corresponding to the customer service work order to be processed, and determining a customer risk label of the user according to the historical behavior data; Judging whether to carry out voiceprint verification on the customer service work order to be processed according to the work order risk grade and the customer risk label; when the judgment result is yes, preprocessing the original call voice corresponding to the customer service work order to be processed to obtain the call voice to be analyzed; Performing voiceprint matching on the conversation voice to be analyzed according to a voiceprint matching model trained in advance to obtain a confidence coefficient corresponding to the conversation voice to be analyzed, wherein the confidence coefficient is used for representing the voiceprint matching degree of the conversation voice to be analyzed and the user; And executing a risk coping process corresponding to the risk control strategy according to the risk control strategy matched with the confidence coefficient.
2. The method according to claim 1, wherein said determining a customer risk tag of said user based on said historical behavior data, in particular comprises: determining corresponding behavior scores of the user on each behavior factor based on at least two preset behavior factors according to the historical behavior data; Carrying out weighted summation according to the behavior scores to obtain risk total scores corresponding to the users; comparing the risk total score with a preset risk threshold, and determining the client risk label as a high risk label when the risk total score is larger than the risk threshold; and when the risk total score does not exceed the risk threshold, determining that the client risk label is a non-high risk label.
3. The method of claim 1, wherein the worksheet risk level comprises high risk, medium risk, and low risk; Judging whether to carry out voiceprint verification on the customer service work order to be processed according to the work order risk grade and the customer risk label, wherein the method specifically comprises the following steps of: if the work order risk level is high risk, judging that voiceprint verification is required to be carried out on the customer service work order to be processed; if the work order risk level is medium risk and the customer risk label is high risk label, judging that voiceprint verification is needed for the customer service work order to be processed; If the work order risk level is low risk or the work order risk level is medium risk and the customer risk label is a non-high risk label, judging that voiceprint verification is not needed for the customer service work order to be processed.
4. The method of claim 1, wherein the preprocessing the original call voice corresponding to the customer service work order to be processed specifically includes: Carrying out channel compensation processing on the original call voice according to an inverse coding reconstruction algorithm, and restoring high-frequency voiceprint characteristics of the call voice to obtain a first call voice; According to a pre-trained voice separation model, performing voice filtering on the first communication voice, and extracting a communication voice fragment corresponding to the user to obtain a second communication voice; And determining emotion parameters of the user according to the second conversation voice, and carrying out normalization processing on the two-way conversation voice according to the emotion parameters to obtain conversation voice to be analyzed.
5. The method of claim 1, wherein the performing voiceprint matching on the call voice to be analyzed according to a pre-trained voiceprint matching model to obtain the confidence corresponding to the call voice to be analyzed specifically comprises: acquiring a first voiceprint feature corresponding to the call voice to be analyzed according to the call voice to be analyzed; determining the feature matching degree of the first voiceprint features and the user pre-stored voiceprint features to obtain a first score; determining a second score according to the voice quality of the call voice to be analyzed; And determining the confidence degree according to the first score and the second score.
6. The method of claim 1, wherein the risk control policies include hierarchical policies corresponding to different confidence intervals; Executing a risk coping process corresponding to the risk control strategy according to the risk control strategy matched with the confidence, wherein the risk coping process comprises the following steps: When the confidence coefficient belongs to a first confidence coefficient interval, executing a processing flow of release business handling; when the confidence coefficient belongs to the second confidence coefficient interval, executing a coping process of sending a risk prompt to the customer service agent; and when the confidence coefficient belongs to the third confidence coefficient interval, executing a coping process of switching to the secondary manual verification.
7. A risk control device based on voiceprint recognition, comprising: The work order risk level determining unit is used for determining the work order risk level corresponding to the customer service work order to be processed according to the predefined business risk rule base; The customer risk tag determining unit is used for acquiring historical behavior data of the user corresponding to the customer service work order to be processed and determining a customer risk tag of the user according to the historical behavior data; The judging unit is used for judging whether to carry out voiceprint verification on the customer service work order to be processed according to the work order risk grade and the customer risk label; the voice optimizing unit is used for preprocessing the original call voice corresponding to the customer service work order to be processed when the judging result is yes, so as to obtain the call voice to be analyzed; The voiceprint matching unit is used for carrying out voiceprint matching on the conversation voice to be analyzed according to a voiceprint matching model trained in advance to obtain a confidence coefficient corresponding to the conversation voice to be analyzed, wherein the confidence coefficient is used for representing the voiceprint matching degree of the conversation voice to be analyzed and the user; and the wind control unit is used for executing a risk coping process corresponding to the risk control strategy according to the risk control strategy matched with the confidence coefficient.
8. A voiceprint recognition based risk control apparatus comprising: Processor, and A memory arranged to store computer executable instructions that, when executed, cause the processor to: Determining a work order risk level corresponding to the customer service work order to be processed according to a predefined business risk rule base; Acquiring historical behavior data of a user corresponding to the customer service work order to be processed, and determining a customer risk label of the user according to the historical behavior data; Judging whether to carry out voiceprint verification on the customer service work order to be processed according to the work order risk grade and the customer risk label; when the judgment result is yes, preprocessing the original call voice corresponding to the customer service work order to be processed to obtain the call voice to be analyzed; Performing voiceprint matching on the conversation voice to be analyzed according to a voiceprint matching model trained in advance to obtain a confidence coefficient corresponding to the conversation voice to be analyzed, wherein the confidence coefficient is used for representing the voiceprint matching degree of the conversation voice to be analyzed and the user; And executing a risk coping process corresponding to the risk control strategy according to the risk control strategy matched with the confidence coefficient.
9. A computer readable storage medium storing one or more programs, which when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the voiceprint recognition based risk control method of any one of claims 1-6.
10. A computer program product comprising a computer program which, when executed by a processor, implements a voiceprint recognition based risk control generating method according to any one of claims 1 to 6.

Description

Risk control method, apparatus, device, storage medium, and computer program product Technical Field The present application relates to the field of wireless communications technologies, and in particular, to a risk control method, apparatus, device, storage medium and computer program product based on voiceprint recognition. Background Voiceprint recognition is widely used in the field of authentication in recent years as a biometric feature recognition technology because of its convenience and non-contact. In the prior art, a typical voiceprint verification scheme generally comprises a flow of receiving a voice command, full-touch voiceprint verification, outputting a binary decision result and the like. However, when the voiceprint verification universal scheme is directly transplanted to the field of telecom customer service, because the specific scene of telecom customer service has the requirements of high risk service, complex call environment and strict user experience, the direct transplantation may have the following technical problems that firstly, because dynamic evaluation on service risk and customer behavior is lacking, indiscriminate full-flow voiceprint verification is directly adopted on all incoming telephone sheets, a large amount of low risk service occupies excessive system resources, and verification efficiency is greatly reduced. And secondly, only a binary result of pass or reject is output, the certainty of identity matching cannot be quantified, security holes are easily caused, and the risk identification accuracy is insufficient. Finally, the general voiceprint model and the processing flow used in the existing voiceprint verification are not optimized for special interferences such as telephone channel compression, cross voice, emotional pronunciation and the like in a telecom customer service scene, so that the recognition accuracy in the scene is poor, and the requirements of actual wind control business on reliability are difficult to meet. Therefore, there is a need for a voiceprint recognition wind control method that can adapt to telecom customer service scenarios. Disclosure of Invention The embodiment of the application provides a risk control method based on voiceprint recognition, which is used for solving the problems that in the prior art, the traditional voiceprint verification scheme is directly applied to a telecom customer service scene, and the system efficiency is low, and the safety and the customer experience are difficult to balance because a full-scale triggering and binary judging mechanism is adopted and the telecom customer service scene interference is not optimized. The embodiment of the application also provides a risk control device based on voiceprint recognition, which is used for solving the problems that in the prior art, the traditional voiceprint verification scheme is directly applied to a telecom customer service scene, and the system efficiency is low, and the safety and the customer experience are difficult to balance because a full triggering and binary judging mechanism is adopted and the telecom customer service scene interference is not optimized. The embodiment of the application also provides risk control equipment based on voiceprint recognition, which is used for solving the problems that in the prior art, the traditional voiceprint verification scheme is directly applied to a telecom customer service scene, and the system efficiency is low, and the safety and the customer experience are difficult to balance because a full-scale triggering and binary judging mechanism is adopted and the disturbance of the telecom customer service scene is not optimized. The embodiment of the application also provides a computer readable storage medium, which is used for solving the problems that in the prior art, the traditional voiceprint verification scheme is directly applied to a telecom customer service scene, and the system efficiency is low, and the safety and the customer experience are difficult to balance because a full triggering and binary judging mechanism is adopted and the telecom customer service scene interference is not optimized. A computer program product is used for solving the problems that in the prior art, the traditional voiceprint verification scheme is directly applied to a telecom customer service scene, and the system efficiency is low, and the safety and the customer experience are difficult to balance due to the adoption of a full triggering and binary judging mechanism and the lack of optimization aiming at the telecom customer service scene interference. The embodiment of the application adopts the following technical scheme: A risk control method based on voiceprint recognition comprises the steps of determining a worksheet risk level corresponding to a customer service worksheet according to a predefined business risk rule base, obtaining historical behavior data of a user corresponding to the customer service wor