WO2019174073A1 - 通话中客户信息修改方法、装置、计算机设备及存储介质 - Google Patents

通话中客户信息修改方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2019174073A1
WO2019174073A1 PCT/CN2018/081509 CN2018081509W WO2019174073A1 WO 2019174073 A1 WO2019174073 A1 WO 2019174073A1 CN 2018081509 W CN2018081509 W CN 2018081509W WO 2019174073 A1 WO2019174073 A1 WO 2019174073A1
Authority
WO
WIPO (PCT)
Prior art keywords
modification
matching degree
feature
target
standard
Prior art date
Application number
PCT/CN2018/081509
Other languages
English (en)
French (fr)
Inventor
张政
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2019174073A1 publication Critical patent/WO2019174073A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/018—Certifying business or products
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/01—Customer relationship services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/01—Customer relationship services
    • G06Q30/012—Providing warranty services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L17/00—Speaker identification or verification
    • G10L17/06—Decision making techniques; Pattern matching strategies

Definitions

  • the present application relates to the field of information management technologies, and in particular, to a method, device, computer device, and storage medium for modifying customer information during a call.
  • the embodiment of the present invention provides a method, a device, a computer device, and a storage medium for modifying customer information during a call, so as to solve the problem that the modification of customer information in the current call cannot guarantee the security of the customer information.
  • the embodiment of the present application provides a method for modifying customer information during a call, including:
  • the modification verification request includes a customer mobile phone number and a feature to be identified
  • the embodiment of the present application provides a device for modifying customer information during a call, including:
  • a modification verification request obtaining module configured to obtain a modification verification request, where the modification verification request includes a customer mobile phone number and a feature to be identified;
  • a standard identification feature acquisition module configured to query a database based on the customer mobile phone number to obtain a corresponding standard identification feature
  • the identification matching processing module is configured to obtain a corresponding recognition matching degree based on the to-be-identified feature and the standard identification feature, and enter the information modification interface if the recognition matching degree is greater than or equal to the preset matching degree;
  • Modifying a data acquisition display module configured to acquire target modification data, display the target modification data in the information modification interface, and set the target modification data to be an unmodifiable state
  • a modification confirmation processing module configured to acquire a modification confirmation instruction, and upload the target modification data to the database.
  • an embodiment of the present application provides a computer device, including a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, where the processor executes the computer The following steps are implemented when reading the instruction:
  • the modification verification request includes a customer mobile phone number and a feature to be identified
  • the embodiment of the present application provides one or more non-volatile readable storage media storing computer readable instructions, when the computer readable instructions are executed by one or more processors, such that the one or Multiple processors perform the following steps:
  • the modification verification request includes a customer mobile phone number and a feature to be identified
  • FIG. 1 is a flowchart of a method for modifying customer information during a call according to Embodiment 1 of the present application;
  • FIG. 2 is a specific flow chart of step S40 of Figure 1;
  • FIG 3 is another specific flow chart of step S40 of Figure 1;
  • FIG. 4 is a schematic diagram of an apparatus for modifying a client information during a call according to Embodiment 2 of the present application;
  • FIG. 5 is a schematic diagram of a computer device in Embodiment 4 of the present application.
  • FIG. 1 is a flow chart showing a method for modifying customer information during a call in the embodiment.
  • the method for modifying customer information in the call is applied to a seat call system provided by a financial institution such as insurance and securities or other institution, and the agent call system includes a server and at least one agent terminal connected to the server, and the agent terminal is a seat person and a client.
  • the terminal used to communicate by telephone.
  • the server is also connected to the client terminal, and the client terminal is a terminal used when the client communicates with the agent.
  • the smart phone is used as the client terminal as an example for description.
  • the method for modifying the customer information in the call is mainly applied to the server of the agent call system, which is used to modify the customer information during the call between the client and the agent, and can ensure the security of the customer information modification.
  • the method for modifying the customer information in the call specifically includes the following steps:
  • S10 Acquire a modification verification request, where the modification verification request includes a customer mobile phone number and a feature to be identified.
  • the modification verification request is a request sent by the smartphone to the server to verify the identity of the client.
  • the customer mobile phone number is the mobile phone number corresponding to the mobile phone card installed on the smart phone, and each customer mobile phone number is associated with the customer information corresponding to a characteristic customer.
  • the modification verification request carries the customer mobile phone number, so as to find and modify the corresponding customer information based on the customer mobile phone number.
  • the feature to be identified is a feature for identifying a customer identity
  • the feature to be identified is a feature acquired when the current verification request is triggered.
  • the to-be-identified feature includes a to-be-recognized character feature and/or a to-be-recognized speech feature for performing identity recognition, wherein the to-be-recognized character feature is a text form that is collected by the smart phone and the customer is based on the security and privacy question of the smart phone prompt.
  • the characteristics of the reply; the voice feature to be recognized is a feature of the voice form recovery that the customer collects based on the security and privacy question of the smart phone prompt.
  • the client when the client sends a modification verification request to the server through the smart phone, the user can carry only the character to be recognized, or only the voice feature to be recognized, or can carry the character to be recognized and the voice feature to be recognized at the same time.
  • the security and confidentiality problem is a problem that the customer pre-sets and stores in the server to protect the security of the customer information.
  • the corresponding modification verification request may be triggered according to the instruction of the agent.
  • the identifier of the security question can be obtained by inputting the identifier **, and the corresponding character to be recognized and/or the feature to be recognized waiting for the recognition feature are input based on the security secret question, and then ## is selected as the to-be-identified
  • the smart phone uploads the to-be-identified feature collected between the two identifiers ** and ## together with the customer mobile phone number as a modification verification request to the server, so as to verify the authenticity of the customer identity. Ensure the security of customer information modification.
  • S20 Query the database based on the customer mobile phone number to obtain a corresponding standard identification feature.
  • the database is connected to the server and is used to store customer information. Each customer information in the database is associated with a corresponding customer mobile phone number to query the corresponding customer information based on the customer mobile phone number.
  • the client can pre-set a number of security issues through the client terminal and store the security secret question in the server, specifically in a database connected to the server.
  • the standard identification feature is a feature set by the customer to verify the identity of the customer when setting security and privacy issues.
  • Each security secret question corresponds to a standard identification feature, and the security secret question and the standard identification feature are stored in a database to be associated with the customer information so that the corresponding standard identification feature can be quickly found based on the customer mobile phone number.
  • the standard identification feature includes a standard text feature and/or a standard phone feature, wherein the standard recognition feature is a preset text form feature corresponding to the security secret question; the standard voice feature is preset to be related to the security secret question. Corresponding features of the phonetic form. It can be understood that when the security problem is set in advance, the customer can configure only standard text features, or only standard voice features, or both standard text features and standard voice features.
  • S30 Acquire a corresponding recognition matching degree based on the feature to be identified and the standard identification feature. If the recognition matching degree is greater than or equal to the preset matching degree, enter the information modification interface.
  • the recognition matching degree is an index for calculating the similarity between the two based on the feature to be identified and the standard identification feature.
  • the preset match is an indicator for evaluating the similarity of the authentication when it passes.
  • the information modification interface is an interface for displaying customer information and allowing modification of customer information.
  • the information modification interface may be an interface displayed on the agent terminal, so that the agent can view the content in the information modification interface and communicate with the client.
  • the identification matching degree is greater than or equal to the preset matching degree, it indicates that the to-be-identified feature in the modification verification request passes the identity verification, and may enter the information modification interface; if the recognition matching degree is less than the preset matching degree, the description is The feature to be identified in the modification verification request does not pass the authentication, and may not enter the information modification interface, so as to prevent the customer whose identity verification fails from modifying the customer information corresponding to the customer mobile phone number, and the security of the customer information may be guaranteed to a certain extent.
  • S40 Obtain the target modification data, display the target modification data in the information modification interface, and set the target modification data to be unmodifiable.
  • the target modification data is data formed by the customer information to be modified.
  • the target modification data may be modified data directly sent by the client to the server through the smart phone, or may be modified data after the error correction by the agent on the modified data directly sent by the smart phone.
  • the target modification data includes a target modification theme and a target modification content corresponding to the target modification theme.
  • the target modification theme is the theme in the customer information to be modified, including but not limited to the mobile phone number, the document number, the frame number, the engine number and the license plate, and may also include the name, gender, age, address, occupation, and office. Address and other information.
  • the target modification content is the specific value or specific content corresponding to the target modification theme.
  • the information modification interface can be accessed to avoid the modification of the customer information during the call, which violates the true intention of the customer, so as to ensure the security of the customer information.
  • the target modification data is displayed on the information modification interface of the agent terminal, so that the agent who is in a call with the client can access the information that the customer needs to modify through the information modification interface, and confirm with the customer by telephone.
  • the target modification data on the information modification interface is set to an unmodifiable state, so that the agent can only view different modifications, thereby preventing the agent from modifying the customer information in violation of regulations to take advantage of the benefits, and further ensuring the security of the customer information.
  • the modification confirmation command is an instruction sent by the customer to the server through the smartphone to confirm that the modification content is correct.
  • the identifier corresponding to the modification confirmation instruction is set in advance, and during the conversation between the agent and the client, the agent may repeat the target modification data displayed on the information modification interface, and prompt the customer to pass the smart phone after confirming the error. Entering an identifier (for example, #**) corresponding to the modification confirmation instruction to the server, so that the server obtains the modification confirmation instruction, so that the target modification data is uploaded to the database and saved in the update database according to the modification confirmation instruction. Stored customer information.
  • the recognition matching degree is obtained based on the to-be-identified feature in the modification verification request and the standard identification feature pre-stored in the database, and the recognition matching degree is compared with the preset matching degree.
  • the recognition matching degree is greater than or equal to the preset matching degree
  • the identity verification is passed to enter the information modification interface to ensure the security of the modification of the customer information during the call.
  • the target modification data is displayed on the information modification interface, and the target modification data is set to an unmodifiable state, so as to prevent the agent from violating the customer information and further protecting the security of the customer information modification.
  • the target modification data may be uploaded to the database after the modification confirmation instruction is obtained, so as to complete the modification of the customer information, and the modification of the customer information may be avoided against the true intention, thereby ensuring the security of the modification of the customer information.
  • the target modification data includes a target modification theme and a corresponding target modification content.
  • the target modification theme is the theme in the customer information to be modified, including but not limited to the mobile phone number, the document number, the frame number, the engine number and the license plate, and may also include the name, gender, age, address, occupation, and office. Address and other information.
  • the target modification content is the specific value or specific content corresponding to the target modification theme.
  • step S30 if the identification matching degree in step S30 is greater than or equal to the preset matching degree, the information modification interface is entered, which specifically includes the following steps:
  • the recognition matching degree is greater than or equal to the preset matching degree
  • the corresponding target security level is obtained based on the identification matching degree and the preset matching degree
  • the information modification interface corresponding to the target security level is entered, and the information modification interface displays and the target security The level corresponding to the modifiable theme and the corresponding original information content.
  • the target security level is a security level when the client identity verification is determined based on the identification matching degree and the preset matching degree.
  • the target security level is specifically any one of a high security level, a medium security level, and a low security level. The higher the security level, the more secure the customer is in authentication.
  • the information modification interface corresponds to the target security level, that is, the three security levels of the high security level, the medium security level, and the low security level respectively correspond to an information modification interface.
  • a modifiable theme corresponding to the target security level and original information content corresponding to the modifiable theme are displayed on each of the information modification interfaces.
  • the modifiable theme is a theme that the customer corresponding to the target security level can modify during the call.
  • the original information content is a specific value or specific content corresponding to the modifiable subject that is pre-stored in the database.
  • the modifiable theme may be one or more of a mobile phone number, a document number, a frame number, an engine number, and a license plate, and may also be information such as name, gender, age, address, occupation, and office address. One or more of them.
  • displaying the target modification data in the information modification interface in step S40 includes the following steps:
  • the target modification theme belongs to the theme in the modifiable theme
  • the original information content and the target modification content corresponding to the target modification theme are displayed on the information modification interface.
  • each information modification interface corresponds to the target security level determined when the authentication is passed, only the modifiable topic corresponding to the target security level is displayed, and therefore, the target modification data acquired by the server can only be modified.
  • Customer information content corresponding to the target security level That is, only the target modification theme in the target modification data belongs to the theme in the modifiable theme, and the target modification content in the target modification data can be modified to replace the original information content corresponding to the modifiable theme, so as to ensure the security of the modification of the customer information. Sex.
  • the comparison display and the target modification theme may be performed on the information modification interface corresponding to the target security level.
  • Corresponding original information content and target modification content enable the agent to view the modified content of the customer information more intuitively, so as to modify and confirm the customer information during the conversation with the customer.
  • the zooming, red highlighting or other highlighting manner may also be floated, and the target modifying theme and its corresponding original information content and target modification content are compared and displayed. Make the agent notice the content to be modified.
  • the corresponding target security level is determined based on the recognition matching degree and the preset matching degree to enter an information modification interface corresponding to the target security level, and the information modification interface is displayed corresponding to the target security level.
  • the modifiable theme and the corresponding original information content can be implemented to determine a modifiable theme that can be modified during the call according to the target security level, so as to ensure the security of the customer information modification.
  • the higher the target security level the more topics it can modify, and the more customer information that can be modified.
  • the target modification theme in the target modification data belongs to the theme in the modifiable theme, so as to ensure that the customer information modification is associated with the target security level, and further ensure the security of the customer information modification.
  • the original information content and the target modification content corresponding to the target modification theme can be compared and displayed, and the process of modifying the customer information can be clearly and intuitively understood, so as to facilitate the telephone confirmation, thereby ensuring that the modification of the customer information reflects the customer. True will.
  • the feature to be identified includes a character feature to be recognized and/or a speech feature to be recognized.
  • the character to be recognized is a feature that the smart phone collects and the user replies in a text form based on the security and privacy question of the smart phone prompt; the voice feature to be recognized is collected by the smart phone and the security secret question of the customer based on the smart phone prompt The characteristics of the voice form reply.
  • Standard recognition features include standard text features and/or standard speech features.
  • the standard recognition feature is a pre-set feature in the form of a text corresponding to the security secret question; the standard speech feature is a preset feature of the voice form corresponding to the security secret question. It can be understood that the character feature to be recognized corresponds to the standard character feature, and the speech feature to be recognized corresponds to the standard phone feature.
  • the preset matching degree includes a first preset matching degree and a second preset matching degree, and the first preset matching degree is greater than the second preset matching degree.
  • the preset matching degree is an index for evaluating the similarity when the authentication is passed, and can be set autonomously according to actual needs, such as setting to 50%, 60% or other values.
  • the two preset matching degrees that is, the first preset matching degree and the second preset matching degree, are configured, and the first preset matching degree is greater than the second preset matching degree (eg, the first preset matching may be performed.
  • the degree is 90%, and the second preset matching degree is 60%), so that the recognition matching degree is respectively compared with the first preset matching degree and the second preset matching degree to determine the target security level when the authentication is passed. .
  • the corresponding matching degree is obtained based on the to-be-identified feature and the standard identification feature in step S30, and specifically includes the following steps:
  • a first recognition matching degree obtained based on the character feature to be recognized and a standard character feature obtained based on the character feature to be recognized and a standard character feature
  • the recognition matching degree is an index for calculating the similarity between the two based on the feature to be identified and the standard identification feature. Since the feature to be identified includes the character feature to be recognized and/or the voice feature to be recognized, the standard recognition feature includes a standard character feature and/or a standard phone feature, and the first recognition matching degree may be obtained according to the character feature to be recognized and the standard character feature; The to-be-recognized speech feature and the standard speech feature acquire a second recognition matching degree, or obtain a second identification matching degree according to the to-be-recognized speech feature and the standard text feature.
  • the second recognition matching degree obtained based on the to-be-recognized speech feature and the standard speech feature or based on the to-be-recognized speech feature and the standard text feature includes the following two cases:
  • the pre-trained voiceprint recognition model is used to extract the voiceprints and the standard voice features respectively, and obtain the voiceprint features and the standard voiceprint features respectively;
  • the cosine similarity algorithm is used to identify the voiceprints.
  • the feature and the standard voiceprint feature perform a cosine similarity calculation to obtain a second recognition match.
  • the voiceprint recognition model is pre-trained and stored in the server based on the voiceprint for speaker identification.
  • the voiceprint recognition model may be a gmm-ubm (Gaussian mixture model-universal background model), or may be trained by other models.
  • the voiceprint recognition model is used to perform voiceprint extraction on the recognized voice feature and the standard voice feature, respectively, and the voiceprint feature to be recognized and the standard voiceprint feature are respectively acquired, so as to be based on the voiceprint feature to be recognized. Identification with standard voiceprint features.
  • the voiceprint recognition model may be used in advance to perform voiceprint extraction on standard voice features, and the obtained standard voiceprint features are stored in a database, and when the identity verification is required, the corresponding standard voiceprint is directly retrieved.
  • the voiceprint recognition model may be used in advance to perform voiceprint extraction on standard voice features, and the obtained standard voiceprint features are stored in a database, and when the identity verification is required, the corresponding standard voiceprint is directly retrieved.
  • Cosine similarity is a measure of the difference between two individuals using the cosine of the angle between two vectors in vector space.
  • the cosine similarity algorithm is an algorithm for calculating the cosine of the angle between two vectors in vector space.
  • a and b are the two vectors of the voiceprint feature to be recognized and the standard voiceprint feature in the vector space
  • x and y are coordinate values of the vector.
  • the voiceprint recognition model is used to perform the voiceprint extraction on the recognized voice feature and the standard voice feature to form the voiceprint feature to be recognized and the standard voiceprint feature in the form of a space vector, and then the cosine similarity is adopted.
  • the algorithm performs a cosine similarity calculation on the voiceprint feature to be identified and the standard voiceprint feature, and uses the obtained cosine value as the second recognition matching degree, so that the acquired second recognition matching degree can well reflect the to-be-identified voice feature and Whether the standard voice feature is the same customer's voice feature for authentication.
  • the preset speech recognition system performs text conversion on the recognized speech features to obtain the character features to be processed; the cosine similarity degree algorithm is used to calculate the cosine similarity calculation for the processed character features and the standard character features to obtain the second recognition. suitability.
  • the speech recognition system is a system that converts speech into words using a built-in speech recognition algorithm.
  • the server may invoke a preset voice recognition system to perform text conversion on the to-be-recognized voice feature to obtain a character to be processed.
  • the preset speech recognition system may adopt an NLP (Natural Language Processing) library to convert the to-be-recognized speech feature into a to-be-processed character.
  • NLP is a natural language processing database of artificial intelligence, which is a database for matching natural language analysis to a machine language that can be recognized by a computer.
  • the NLP library can use the Stanford CoreNLP developed by the Stanford Natural Speech Processing Team, and can quickly process the recognized speech features for parsing to obtain the features of the text to be processed.
  • the cosine similarity calculation using the cosine similarity algorithm to calculate the cosine similarity calculation includes the following steps: (1) performing word segmentation processing on the processed character feature and the standard text feature, such as ABCDECF after the character feature segmentation to be processed. Seven words, the standard text feature segmentation includes the eight words ABCGHEIF. (2) List all words appearing in the pending text features and standard text features, including the nine words ABCDEFGHI. (3) Calculate the word frequency of the words appearing in the character of the character to be processed and the standard character, respectively, and the word frequency of the character to be processed is A1, B1, C2, D1, E1, F1, G0, H0, I0; The word frequency is A1, B1, C1, D0, E1, F1, H1, G1, I1.
  • the word frequency vector of the character feature to be processed is (1,1,2,1,1,1,0,0,0), the standard text feature and The word frequency vector is (1,1,1,0,1,1,1,1).
  • the cosine similarity algorithm is used to calculate the cosine similarity. The calculation process is as follows:
  • the preset speech recognition system is used to perform character conversion on the recognized speech features to obtain the character features to be processed; then, the word features, the summary words, the word frequency calculations, and the word frequency are respectively processed.
  • the vector is used to calculate the cosine similarity of the word frequency vector converted by the cosine similarity algorithm, and the obtained cosine value is used as the second identification matching degree, so that the acquired second identification matching degree can clearly reflect the customer based security. Whether the speech feature to be recognized by the confidentiality question matches the pre-stored standard text feature for identity verification.
  • the information modification interface is entered, which specifically includes the following situations:
  • the information modification interface corresponding to the high security level is entered.
  • the first preset matching degree is greater than the second preset matching degree (eg, the first preset matching degree is 90%, and the second preset matching degree is 60%), if the feature is identified based on the feature to be identified and the standard
  • the recognition matching degree includes the first identification matching degree and the second identification matching degree
  • the first identification matching degree ⁇ the first preset matching degree, and the second identification matching degree ⁇ the first preset matching degree the identity verification is performed
  • the target security level when passing is a high security level.
  • the information modification interface corresponding to the high security level can be entered, and a large number of modifiable topics and their corresponding original information contents can be displayed on the information modification interface.
  • the information modification interface corresponding to the low security level is entered.
  • the first preset matching degree is greater than the second preset matching degree (eg, the first preset matching degree is 90%, and the second preset matching degree is 60%), if the feature is identified based on the feature to be identified and the standard
  • the acquired matching degree is between the first preset matching degree and the second preset matching degree, it indicates that although the identity is verified, the target security level is low, which is a low security level.
  • the degree ⁇ the second preset matching degree the target security level when the authentication is passed is a low security level, and at this time, the information modification interface corresponding to the low security level can be entered, and the number of the information modification interface can be displayed on the information modification interface.
  • the subject and its corresponding original information content can be modified.
  • any one of the first identification matching degree and the second identification matching degree is greater than or equal to the first preset matching degree, the other is smaller than the first preset matching degree and greater than or equal to the second preset matching degree. , enter the information modification interface corresponding to the middle security level.
  • the medium security level is a level of security between a high security level and a low security level.
  • the number of modifiable topics displayed on the information modification interface corresponding to the medium security level is also between the number of modifiable topics displayed on the information modification interface corresponding to the high security level and the low security level.
  • the first preset matching degree is greater than the second preset matching degree (eg, the first preset matching degree is 90%, and the second preset matching degree is 60%), and the identification matching degree includes the first identification matching degree.
  • the target security level when the authentication is passed can be determined as the medium security level: first, if the first preset matching degree > the first identification matching degree ⁇ the second pre- The matching degree is set, and the second identification matching degree is ⁇ the first preset matching degree; the second is that the first preset matching degree > the second identification matching degree ⁇ the second preset matching degree, and the first identification matching degree ⁇ A preset match.
  • the identity verification fails and the information modification interface cannot be accessed.
  • the recognition matching degree is an index for calculating the similarity between the two based on the feature to be identified and the standard identification feature
  • the second preset matching degree is a recognition matching degree with a small value set in advance, which is used for verifying the identity.
  • the minimum indicator therefore, the first identification matching degree and the second identification matching degree both need to be greater than or equal to the second preset matching degree. Specifically, if the first recognition matching degree is smaller than the second preset matching degree, or the second identification matching degree is smaller than the second preset matching degree, or the first identification matching degree and the second identification matching degree are simultaneously smaller than the second preset
  • the matching degree will cause the authentication to fail. At this time, the information modification interface cannot be entered to ensure the security of the customer information.
  • the first recognition matching degree is first obtained based on the character feature to be recognized and the standard character feature
  • the second recognition matching degree is obtained based on the to-be-recognized speech feature and the standard speech feature or based on the to-be-recognized speech feature and the standard character feature. And determining a target security level when the authentication is passed based on the first identification matching degree and the second identification matching degree.
  • the target security level corresponding to the to-be-identified feature and the standard identification feature is a high security level
  • Any one of the medium security level and the low security level respectively enters an information modification interface corresponding to the target security level, so that the modification of the customer information is associated with the security level of the identity verification to ensure the security of the modification of the customer information.
  • the acquiring target modification data in step S40 specifically includes the following steps:
  • S411 Acquire voice modification data that is uploaded twice in succession.
  • the voice modification data is data that the client uploads to the server in a voice form to reflect the customer information to be modified through the smart phone, that is, the data to be modified that is uploaded by the customer and expressed in a voice form.
  • voice modification data In order to avoid uploading voice modification data due to false triggering, or to avoid uploading incorrect voice modification data, it is necessary to obtain voice modification data that is uploaded twice in succession. If the voice modification data uploaded twice consecutively is the same data, it may be explained
  • the data is modified according to the voice uploaded by the customer subjectively.
  • a start identifier such as #*#
  • an end identifier such as ###
  • the time interval of the voice modification data that is uploaded twice consecutively may be limited to a preset time, to avoid uploading another voice modification data after a long time interval after the first time the voice modification data is uploaded, thereby causing the system to process the process or The thread is always waiting, affecting the normal operation of the server.
  • the preset time is a preset time interval for determining whether the voice modification data uploaded twice consecutively is valid data. That is, if the time interval between the two uploaded voice modification data is too long, which is greater than the preset time, the two are not considered to be consecutive uploads as described in the embodiment.
  • S412 Perform text conversion and keyword extraction on the voice modification data, and obtain a modified theme keyword and a modified content keyword.
  • the preset voice recognition system (NLP mentioned above) is called to perform text conversion on the voice modification data to convert the voice modification data in the voice form into the original modification data in the text form.
  • a preset keyword extraction algorithm including but not limited to the TextRank algorithm and the TF-IDF algorithm, etc.
  • the original modified data in the text form is subjected to keyword extraction to obtain the modified theme keyword and the modified content keyword.
  • the voice modification data uploaded by the customer is converted into the original modified data as follows: I want to modify the mobile phone number to XXX, and the modified theme keyword extracted is the mobile phone number, and the modified content keyword is XXX.
  • the first modifiable threshold is a pre-set threshold associated with the modified topic, and is a threshold for evaluating whether the modified subject can be determined.
  • the cosine similarity degree algorithm is used to calculate the cosine similarity of the modified topic keywords in the two successively uploaded voice modification data to obtain the similarity between the two, and the calculation process is performed on the text feature and the standard text feature as described above. The process of cosine similarity calculation is consistent. To avoid redundancy, the calculation process will not be detailed.
  • the first modifiable threshold may be set to 70%. If the similarity of the modified topic keyword in the voice modification data uploaded twice reaches the first modifiable threshold, the corresponding keyword is obtained based on the modified topic keyword. Goal to modify the theme. For example, the similarity between the two consecutively uploaded modified topic keywords of the mobile phone number and the mobile phone number is 75%, which is greater than the first modifiable threshold value of 70%. At this time, the database may be searched based on the mobile phone number or the mobile phone number to determine corresponding thereto. The modifiable theme in the customer information, as the target to modify the theme. For another example, if the similarity of the two consecutively uploaded modified topic keywords of the mobile phone number and the document number is 33.3%, which is smaller than the first modifiable threshold, the corresponding target modification theme cannot be obtained based on the modified topic keyword.
  • the second modifiable threshold is a preset threshold associated with the modified content, and is a threshold for evaluating whether the modified content can be determined.
  • the second modifiable threshold may or may not coincide with the first modifiable threshold, and is determined according to the content to be modified.
  • the cosine similarity degree algorithm is used to calculate the cosine similarity of the modified content keywords in the two consecutively uploaded voice modification data to obtain the similarity between the two. To avoid redundancy, the calculation process is not detailed one by one. Since the modified content involves the substantive content of the customer information, it is generally required that the second modifiable threshold setting is large, even close to 100%, especially when the mobile phone number, the document number, the frame number, the engine number and the license plate are involved.
  • S415 Modify the content and target modification content based on the target, and obtain the target modification data.
  • the target modification data can be acquired based on the target modification theme and the target modification content only after the target modification theme is acquired according to step S413 and the target modification content is acquired according to step S414. That is, if the similarity of the modified topic keyword in the voice modification data uploaded twice fails to reach the first modifiable threshold, or the similarity of the modified content keyword in the voice modification data uploaded twice consecutively does not reach the second If the threshold is modified, the target modification data is not obtained, and steps S411-S415 are performed again, or the target modification data is determined by other methods.
  • the voice modification data that is uploaded twice consecutively is obtained to avoid false triggering uploading or uploading errors, thereby ensuring that the modification of the customer information conforms to the true intention of the customer and ensures the security of the customer information.
  • the modified theme keyword and the modified content keyword are obtained; and the similarity is calculated based on the modified theme keywords obtained twice consecutively, and the similarity is first
  • the threshold may be modified for comparison to determine the target modification theme; the similarity is calculated based on the modified content keywords obtained twice consecutively, and the similarity is compared with the second modifiable threshold to obtain the target modification content.
  • the target modification theme and the target modification content are determined at the same time, the corresponding target modification data can be obtained to ensure the authenticity of the target modification data and ensure the security of the customer information modification.
  • the similarity of the modified topic keywords in the voice modification data does not reach the first modifiable threshold, or the similarity of the modified content keywords in the voice modification data does not reach the second To modify the threshold, you need to perform steps S411-S415 to perform repeated verification.
  • the number of repeated verifications reaches the preset number of times (pre-autonomously set the number of times, which can be 2 times, 3 times, or other times)
  • the error can be modified by the agent.
  • the acquisition target modification data in step S40 specifically includes the following steps:
  • S421 Acquire error modification data, where the error modification data includes an error modification theme and a corresponding error modification content.
  • the error modification data is data sent by the agent to the server through the agent terminal to reflect the customer information to be modified.
  • the error modification theme is that the agent determines the topic in the customer information to be modified according to the voice modification data uploaded by the client in the call with the client.
  • the error modification content is a specific value or specific content corresponding to the error modification theme.
  • S422 Acquire a corresponding modification restriction rule and original modification content based on the error modification theme.
  • the modification restriction rule corresponds to the error modification topic and is used to restrict the rules for the agent to modify the data. Modifying the restriction rules can be limited based on the number of modified digits, similar pinyin and fuzzy tones.
  • the modification restriction rule based on the modified number of bits can be set as follows: when the mobile phone number is modified, the number of modified digits is controlled to 2 digits; or when the frame number is modified, the number of modified digits is controlled to 5 digits.
  • the number of modified bits can be set autonomously by the number of bits of the value corresponding to the error modification topic.
  • a modification restriction rule based on similar pinyin can be used to implement the same text replacement of Pinyin.
  • the fuzzy tone-based modification restriction rule can be used to implement the equipotential replacement of the fuzzy sounds such as the front and rear nasal sounds and the squeaky sound.
  • the original modified content is the modified content corresponding to the error modification theme determined by the customer using steps S411-S415, and the original modified content may be one or more, that is, the modification identified by the customer in the voice modification data uploaded multiple times. Content, but these modifications are not the same in two consecutive uploads.
  • the modification restriction rule determined in step S422 is used for verification, and the error modification content at the time of verification is used as the target modification content. If the modification restriction rule corresponding to the error modification topic is a modification rule based on the number of bits, it is judged whether the difference between the error modification content and the original modification content is within the preset modification number, and if so, the test passes and the error is Modify the content as the target to modify the content.
  • steps S421-S423 during the process of modifying the customer information in the call with the agent, the customer may perform the assist modification by using the step S411-SS415 because the pronunciation or other reasons may cause the customer to perform the modification.
  • the customer information is modified; and the modification restriction rule is set to restrict the agent from modifying the customer information to prevent malicious modification.
  • FIG. 4 is a block diagram showing the principle of the in-call client information modification apparatus corresponding to the one-to-one client information modification method in the first embodiment.
  • the in-call customer information modification apparatus includes a modification verification request acquisition module 10, a standard identification feature acquisition module 20, an identification matching processing module 30, a modification data acquisition display module 40, and a modification confirmation processing module 50.
  • the implementation functions of the modification verification request acquisition module 10, the standard identification feature acquisition module 20, the identification matching processing module 30, the modification data acquisition display module 40, and the modification confirmation processing module 50 correspond to the in-call client information modification method in Embodiment 1.
  • the steps are one-to-one correspondence, and the details are not described in detail in order to avoid redundancy.
  • the modification verification request obtaining module 10 is configured to obtain a modification verification request, where the modification verification request includes a customer mobile phone number and a feature to be identified.
  • the standard identification feature acquisition module 20 is configured to query a database based on the customer mobile phone number to obtain a corresponding standard identification feature.
  • the identification matching processing module 30 is configured to obtain a corresponding recognition matching degree based on the to-be-identified feature and the standard identification feature, and enter the information modification interface if the recognition matching degree is greater than or equal to the preset matching degree.
  • the modified data acquisition display module 40 is configured to acquire target modification data, display target modification data in the information modification interface, and set the target modification data to be unmodifiable.
  • the modification confirmation processing module 50 is configured to obtain a modification confirmation instruction, and upload the target modification data to the database.
  • the target modification data includes a target modification theme and a corresponding target modification content.
  • the identification matching processing module 30 is configured to obtain a corresponding target security level based on the recognition matching degree and the preset matching degree, and enter an information modification interface corresponding to the target security level, if the matching degree is greater than or equal to the preset matching degree, The modifiable theme corresponding to the target security level and the corresponding original information content are displayed on the information modification interface.
  • the modified data acquisition display module 40 is configured to compare the original information content and the target modification content corresponding to the target modification theme on the information modification interface if the target modification theme belongs to the theme in the modifiable theme.
  • the feature to be identified comprises a character feature to be recognized and/or a speech feature to be recognized.
  • Standard recognition features include standard text features and/or standard speech features.
  • the preset matching degree includes a first preset matching degree and a second preset matching degree, and the first preset matching degree is greater than the second preset matching degree.
  • the target security level is specifically one of a high security level, a medium security level, and a low security level.
  • the identification matching processing module 30 includes an identification matching degree obtaining unit 311 for first recognition matching degree acquired based on the character feature to be recognized and the standard character feature; and/or based on the to-be-recognized speech feature and the standard speech feature or based on the to-be-recognized speech The second recognition match obtained by the feature and the standard text feature.
  • the identification matching processing module 30 further includes a first matching processing unit 321, a second matching processing unit 322, and a third matching processing unit 323.
  • the first matching processing unit 321 is configured to enter an information modification interface corresponding to the high security level if both the first identification matching degree and the second identification matching degree are greater than or equal to the first preset matching degree.
  • the second matching processing unit 322 is configured to: if the first identification matching degree and the second identification matching degree are both smaller than the first preset matching degree and both are greater than or equal to the second preset matching degree, enter the information modification corresponding to the low security level. interface.
  • the third matching processing unit 323 is configured to: if any one of the first identification matching degree and the second identification matching degree is greater than or equal to the first preset matching degree, and the other is smaller than the first preset matching degree and greater than or equal to the second If the preset matching degree is entered, the information modification interface corresponding to the middle security level is entered.
  • the recognition matching degree acquisition unit 311 includes a voiceprint matching processing sub-unit 3111 or a voice matching processing sub-unit 3112.
  • the voiceprint matching processing sub-unit 3111 is configured to perform voiceprint extraction on the recognized voice feature and the standard voice feature respectively by using the pre-trained voiceprint recognition model, respectively acquiring the voiceprint feature to be recognized and the standard voiceprint feature; using cosine similarity
  • the algorithm performs a cosine similarity calculation on the identified voiceprint feature and the standard voiceprint feature to obtain a second recognition match degree.
  • the speech matching processing sub-unit 3112 is configured to perform text conversion on the speech feature to be recognized by using a preset speech recognition system to obtain a character to be processed; and use a cosine similarity algorithm to perform cosine similarity calculation on the processed character feature and the standard character feature, Obtain a second identification match.
  • the modified data acquisition display module 40 includes a voice modification data acquisition unit 411, a keyword extraction acquisition unit 412, a target modification theme acquisition unit 413, a target modification content acquisition unit 414, and a target modification data acquisition unit 415.
  • the voice modification data acquiring unit 411 is configured to acquire voice modification data that is uploaded twice.
  • the keyword extraction obtaining unit 412 is configured to perform text conversion and keyword extraction on the voice modification data, and obtain a modified theme keyword and a modified content keyword.
  • the target modification theme acquiring unit 413 is configured to obtain a corresponding target modification theme based on the modified topic keyword if the similarity of the modified topic keyword in the voice modification data uploaded twice consecutively reaches the first modifiable threshold.
  • the target modification content obtaining unit 414 is configured to obtain the corresponding target modification content based on the modified content keyword if the similarity of the modified content keyword in the voice modification data uploaded twice consecutively reaches the second modifiable threshold.
  • the target modification data obtaining unit 415 is configured to modify the content based on the target modification target and the target modification data.
  • the modified data acquisition display module 40 includes an error modification data acquisition unit 421, a restriction rule and content acquisition unit 422, and a modification verification processing unit 423.
  • the error modification data acquisition unit 421 is configured to acquire error modification data, where the error modification data includes an error modification topic and a corresponding error modification content.
  • the restriction rule and content acquisition unit 422 is configured to acquire the corresponding modification restriction rule and the original modification content based on the error modification theme.
  • the modification verification processing unit 423 is configured to perform the verification processing on the original modification content and the error modification content by using the modification restriction rule, and if the verification passes, the error modification content is used as the target modification content.
  • This embodiment provides one or more non-volatile readable storage media having computer readable instructions stored thereon.
  • the one or more non-transitory readable storage mediums storing computer readable instructions that, when executed by one or more processors, cause one or more processors to perform the call in embodiment 1
  • the method of modifying customer information in order to avoid duplication is not repeated here.
  • the computer readable instructions are executed by the processor, the functions of the modules/units in the modification of the client information in the call in Embodiment 2 are implemented. To avoid repetition, details are not described herein again.
  • non-volatile readable storage media storing computer readable instructions may comprise: any entity or device capable of carrying the computer readable instructions, a recording medium, a USB flash drive, a mobile hard drive, a magnetic Discs, optical discs, computer memories, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier signals, and telecommunications signals.
  • FIG. 5 is a schematic diagram of a computer device according to an embodiment of the present application.
  • computer device 60 of this embodiment includes a processor 61, a memory 62, and computer readable instructions 63 stored in memory 62 and operative on processor 61.
  • the processor 61 implements the steps of the in-call client information modification method in the above-described embodiment 1 when the computer readable instructions 63 are executed, such as steps S10-S50 shown in FIG.
  • the processor 61 executes the computer readable instructions 63
  • the functions of the module/unit of the in-call client information modification apparatus in the second embodiment are implemented, for example, the modification verification request acquisition module 10 and the standard identification feature acquisition module shown in FIG. 20. Identify the functions of the matching processing module 30, the modified data acquisition display module 40, and the modification confirmation processing module 50.
  • computer readable instructions 63 may be partitioned into one or more modules/units, one or more modules/units being stored in memory 62 and executed by processor 61 to complete the application.
  • the one or more modules/units can be an instruction segment of a series of computer readable instructions capable of performing a particular function for describing the execution of computer readable instructions 63 in computer device 60.
  • the computer readable instructions 63 may be divided into a modified verification request acquisition module 10, a standard identification feature acquisition module 20, an identification matching processing module 30, a modified data acquisition display module 40, and a modification confirmation processing module 50, each module having specific functions such as implementation As described in Example 2, it will not be repeated here.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

Abstract

本申请公开了一种通话中客户信息修改方法、装置、计算机设备及存储介质。该通话中客户信息修改方法包括：获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；基于所述客户手机号查询数据库，获取对应的标准识别特征；基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；获取修改确认指令，将所述目标修改数据上传到所述数据库。该通话中客户信息修改方法需在基于识别匹配度与预设匹配度进行身份验证通过后才可进行客户信息修改，可以保证客户信息修改的安全。

Description

通话中客户信息修改方法、装置、计算机设备及存储介质
本申请以2018年03月12日提交的申请号为201810198383.6，名称为“通话中客户信息修改方法、装置、计算机设备及存储介质”的中国发明专利申请为基础，并要求其优先权。
技术领域
本申请涉及信息管理技术领域，尤其涉及一种通话中客户信息修改方法、装置、计算机设备及存储介质。
背景技术
随着通信技术的发展，企业办公越来越呈现电子化和远程化，使得企业之间或者企业和客户之间通过电话沟通成为常态。在保险和证券等金融服务行业中，通常会配备坐席团队与客户进行电话沟通，以便给客户提供更贴心的服务。当前客户与坐席人员的电话沟通一般限于业务咨询等简单业务，无法实现在通话过程中对客户信息进行修改这一项业务，其原因在于，当前客户信息修改这一项业务通常需客户本人到柜台进行面对面身份验证后才可进行，以保证客户信息的安全。而在通话过程中对客户信息进行修改可能存在坐席人员违规修改客户信息以套取利益的现象，使得其修改的客户信息无法识别是否为客户的真实需求，影响客户的利益。
发明内容
本申请实施例提供一种通话中客户信息修改方法、装置、计算机设备及存储介质，以解决当前通话中修改客户信息无法保障客户信息安全的问题。
第一方面，本申请实施例提供一种通话中客户信息修改方法，包括：
获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
基于所述客户手机号查询数据库，获取对应的标准识别特征；
基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
获取修改确认指令，将所述目标修改数据上传到所述数据库。
第二方面，本申请实施例提供一种通话中客户信息修改装置，包括：
修改验证请求获取模块，用于获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
标准识别特征获取模块，用于基于所述客户手机号查询数据库，获取对应的标准识别特征；
识别匹配处理模块，用于基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
修改数据获取显示模块，用于获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
修改确认处理模块，用于获取修改确认指令，将所述目标修改数据上传到所述数据库。
第三方面，本申请实施例提供一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现如下步骤：
获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
基于所述客户手机号查询数据库，获取对应的标准识别特征；
基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
获取修改确认指令，将所述目标修改数据上传到所述数据库。
第四方面，本申请实施例提供一个或多个存储有计算机可读指令的非易失性可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤：
获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
基于所述客户手机号查询数据库，获取对应的标准识别特征；
基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
获取修改确认指令，将所述目标修改数据上传到所述数据库。
本申请的一个或多个实施例的细节在下面的附图及描述中提出。本申请的其他特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例的描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例1中通话中客户信息修改方法的一流程图；
图2是图1中步骤S40的一具体流程图；
图3是图1中步骤S40的另一具体流程图；
图4是本申请实施例2中通话中客户信息修改方法装置的一示意图；
图5是本申请实施例4中计算机设备的一示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
实施例1
图1示出本实施例中通话中客户信息修改方法的流程图。该通话中客户信息修改方法应用在保险和证券等金融机构或者其他机构所配备的坐席通话系统中，该坐席通话系统包括服务器和与服务器相连的至少一个坐席终端，该坐席终端为坐席人员与客户进行电话沟 通时所采用的终端。该服务器还与客户终端通信相连，该客户终端为客户与坐席人员进行电话沟通时所采用的终端，本实施例中以智能手机作为客户终端为例进行说明。该通话中客户信息修改方法主要应用在坐席通话系统的服务器上，用于实现在客户与坐席人员通话过程中对客户信息进行修改，并可保证客户信息修改的安全性。如图1所示，该通话中客户信息修改方法具体包括如下步骤：
S10：获取修改验证请求，修改验证请求包括客户手机号和待识别特征。
修改验证请求是智能手机给服务器发送的用于验证客户身份的请求。客户手机号是智能手机上装配的手机卡对应的手机号，每一客户手机号关联一特征客户对应的客户信息。客户采用智能手机给服务器发送修改验证请求时，修改验证请求中携带有客户手机号，以便基于该客户手机号查找和修改其对应的客户信息。
待识别特征是用于识别客户身份的特征，该待识别特征是在触发本次修改验证请求时采集到的特征。具体地，待识别特征包括用于进行身份识别的待识别文字特征和/或待识别语音特征，其中，待识别文字特征是智能手机采集到的、客户基于智能手机提示的安全保密问题进行文字形式回复的特征；待识别语音特征是智能手机采集到的、客户基于智能手机提示的安全保密问题进行语音形式回复的特征。可以理解地，客户通过智能手机向服务器发送修改验证请求时，可以仅携带待识别文字特征，也可以仅携带待识别语音特征，或者可以同时携带待识别文字特征和待识别语音特征。其中，安全保密问题是客户预先设置并存储在服务器中的、用于保障客户信息安全的问题。
本实施例中，当客户通过智能手机与坐席人员通过坐席终端进行通话时，可以根据坐席人员的指引触发输入相应的修改验证请求。例如，可以通过输入**这一标识符以获取安全保密问题的提示，并基于该安全保密问题输入相应的待识别文字特征和/或待识别语音特征等待识别特征，再输入##作为待识别特征输入结束的标识符，智能手机会将**和##这两个标识符之间采集到的待识别特征和客户手机号一起作为修改验证请求上传给服务器，以便验证客户身份的真实性，保证客户信息修改的安全性。
S20：基于客户手机号查询数据库，获取对应的标准识别特征。
数据库与服务器相连，用于存储客户信息。数据库中的每一条客户信息与对应的客户手机号关联，以便基于客户手机号进行查询到对应的客户信息。如上所述，客户可通过客户终端预先设置若干安全保密问题，并将该安全保密问题存储在服务器中，具体存储在与服务器相连的数据库中。标准识别特征是客户在设置安全保密问题时一并设置的用于验证客户身份的特征。每一安全保密问题对应一标准识别特征，并将该安全保密问题和标准识别特征存储在数据库中，使其与客户信息关联，以便基于客户手机号可快速查找到相应的标准识别特征。
具体地，标准识别特征包括标准文字特征和/或标准语音特征，其中，标准识别特征是预先设置的与安全保密问题相对应的文字形式的特征；标准语音特征是预先设置的与安全保密问题相对应的语音形式的特征。可以理解地，客户在预先设置安全保密问题时，可以仅配置标准文字特征，也可以仅配置标准语音特征，或者可以同时配置标准文字特征和标准语音特征。
S30：基于待识别特征和标准识别特征获取对应的识别匹配度，若识别匹配度大于或等于预设匹配度，则进入信息修改界面。
识别匹配度是基于待识别特征和标准识别特征进行计算获取的用于体现两者相似度的指标。预设匹配度是用于评价身份验证通过时其相似度的指标。信息修改界面是用于显 示客户信息并允许进行客户信息修改的界面。该信息修改界面可以是坐席终端上显示的界面，以使坐席人员可以查看信息修改界面中的内容，与客户进行沟通。
本实施例中，若识别匹配度大于或等于预设匹配度，则说明该修改验证请求中的待识别特征通过身份验证，可以进入信息修改界面；若识别匹配度小于预设匹配度，则说明该修改验证请求中的待识别特征未通过身份验证，不可以进入信息修改界面，以避免身份验证未通过的客户修改客户手机号对应的客户信息，可以在一定程度上保障客户信息的安全。
S40：获取目标修改数据，在信息修改界面中显示目标修改数据，设置目标修改数据为不可修改状态。
目标修改数据为所要修改的客户信息形成的数据。该目标修改数据可以是客户通过智能手机向服务器直接发送的修改数据，也可以是坐席人员对智能手机直接发送的修改数据进行误差修正后的修改数据。目标修改数据包括目标修改主题和与目标修改主题相对应的目标修改内容。其中，目标修改主题是所要修改的客户信息中的主题，包括但不限于手机号、证件号、车架号、发动机号和车牌等信息，还可以包括姓名、性别、年龄、住址、职业和办公地址等信息。目标修改内容是目标修改主题对应的具体值或具体内容。
本实施例中，在基于待识别特征和标准识别特征进行身份验证通过后，才可进入信息修改界面，以避免通话过程中客户信息的修改违背客户的真实意愿，以保障客户信息的安全。在坐席终端的信息修改界面上显示目标修改数据，可以使与客户进行通话的坐席人员可以通过信息修改界面查阅到客户所需要修改的信息，并与客户进行电话确认。具体地，将信息修改界面上的目标修改数据设置为不可修改状态，使得坐席人员只能查看不同修改，从而避免坐席人员违规修改客户信息以套取利益，进一步保障客户信息的安全。
S50：获取修改确认指令，将目标修改数据上传到数据库。
修改确认指令是客户通过智能手机给服务器发送的用于确认修改内容无误的指令。本实施例中，预先设置好修改确认指令对应的标识符，在坐席人员与客户通话过程中，可由坐席人员复述信息修改界面上显示的目标修改数据，并提示客户在确认无误后可通过智能手机向服务器输入与修改确认指令对应的标识符(例如，#**)，以使服务器获取到该修改确认指令，以使基于该修改确认指令将目标修改数据上传到数据库中保存，覆盖更新数据库中存储的客户信息。
本实施例所提供的通话中客户信息修改方法中，基于修改验证请求中的待识别特征与预先存储在数据库中的标准识别特征获取识别匹配度，将该识别匹配度与预设匹配度进行比较，在识别匹配度大于或等于预设匹配度时，身份验证通过，以进入信息修改界面，以保障通话中客户信息修改的安全。在信息修改界面上显示目标修改数据，并设置该目标修改数据为不可修改状态，以避免坐席人员违规修改客户信息，进一步保障客户信息修改的安全。并且，需在获取修改确认指令之后才可将目标修改数据上传数据库，以完成客户信息修改，可避免客户信息修改违背其真实意愿，从而保障客户信息修改的安全。
在一具体实施方式中，目标修改数据包括目标修改主题和对应的目标修改内容。其中，目标修改主题是所要修改的客户信息中的主题，包括但不限于手机号、证件号、车架号、发动机号和车牌等信息，还可以包括姓名、性别、年龄、住址、职业和办公地址等信息。目标修改内容是目标修改主题对应的具体值或具体内容。
在该具体实施方式中，步骤S30中的若识别匹配度大于或等于预设匹配度，则进入信息修改界面，具体包括如下步骤：
若识别匹配度大于或等于预设匹配度，则基于识别匹配度和预设匹配度获取对应的目标安全等级，并进入与目标安全等级相对应的信息修改界面，信息修改界面上显示与目标安全等级相对应的可修改主题和对应的原始信息内容。
其中，目标安全等级是基于识别匹配度和预设匹配度确定的该客户身份验证通过时的安全等级。本实施例中，目标安全等级具体为高安全等级、中安全等级和低安全等级中的任一个。安全等级越高，说明该客户在身份验证时越安全。
具体地，信息修改界面与目标安全等级相对应，即高安全等级、中安全等级和低安全等级这三个安全等级分别对应一信息修改界面。在每一信息修改界面上显示与目标安全等级相对应的可修改主题和与可修改主题对应的原始信息内容。其中，可修改主题是与该目标安全等级相对应的客户可在通话过程中修改的主题。原始信息内容是预先存储在数据库中的与该可修改主题相对应的具体值或具体内容。本实施例中，可修改主题可以是手机号、证件号、车架号、发动机号和车牌等信息中的一个或多个，还可以是姓名、性别、年龄、住址、职业和办公地址等信息中的一个或多个。
在该具体实施方式中，步骤S40中的在信息修改界面中显示目标修改数据，具体包括如下步骤：
若目标修改主题属于可修改主题中的主题时，则在信息修改界面上对比显示与目标修改主题相对应的原始信息内容和目标修改内容。
由于每一信息修改界面与身份验证通过时确定的目标安全等级相对应，其上只显示与该目标安全等级相对应的可修改主题，因此，服务器获取到的目标修改数据也只能修改与该目标安全等级相对应的客户信息内容。即只有目标修改数据中的目标修改主题属于可修改主题中的主题时，才可将目标修改数据中的目标修改内容修改替换与可修改主题相对应的原始信息内容，以保证客户信息修改的安全性。
进一步地，为了更直观地查看和了解客户信息的修改过程，在目标修改主题属于可修改主题中的主题时，可在与目标安全等级相对应的信息修改界面上，对比显示与目标修改主题相对应的原始信息内容和目标修改内容，使得坐席人员可更直观地查看到客户信息修改的内容，以便于在与客户通话过程中进行客户信息修改确认。
进一步地，在信息修改界面上显示的所有可修改主题的基础上，还可以浮动放大、标红或其他突出显示方式，对比显示该目标修改主题及其对应的原始信息内容和目标修改内容，以使坐席人员注意到所要修改的内容。
在该具体实施方式中，基于识别匹配度和预设匹配度确定其对应的目标安全等级，以进入与目标安全等级相对应的信息修改界面，在该信息修改界面上显示与其目标安全等级相对应的可修改主题和对应的原始信息内容，可实现根据目标安全等级确定可在通话过程中进行修改的可修改主题，以保证客户信息修改的安全。一般而言，目标安全等级越高，其可修改主题越多，能够修改的客户信息越多。在获取目标修改数据时，需判断目标修改数据中的目标修改主题是否属于可修改主题中的主题，以保证客户信息修改与其目标安全等级相关联，进一步保证客户信息修改的安全。在信息修改界面上还可对比显示与目标修改主题相对应的原始信息内容和目标修改内容，可以清楚直观地了解客户信息修改的过程中，以便于进行电话确认，从而确保客户信息的修改反映客户的真实意愿。
在一具体实施方式中，待识别特征包括待识别文字特征和/或待识别语音特征。其中，待识别文字特征是智能手机采集到的、客户基于智能手机提示的安全保密问题进行文字形式回复的特征；待识别语音特征是智能手机采集到的、客户基于智能手机提示的安全保密 问题进行语音形式回复的特征。
标准识别特征包括标准文字特征和/或标准语音特征。其中，标准识别特征是预先设置的与安全保密问题相对应的文字形式的特征；标准语音特征是预先设置的与安全保密问题相对应的语音形式的特征。可以理解地，待识别文字特征与标准文字特征相对应，待识别语音特征与标准语音特征相对应。
预设匹配度包括第一预设匹配度和第二预设匹配度，第一预设匹配度大于第二预设匹配度。预设匹配度是用于评价身份验证通过时其相似度的指标，可以根据实际需求自主设置，如设置为50％，60％或其他数值。本实施例中，配置两个预设匹配度，即第一预设匹配度和第二预设匹配度，且第一预设匹配度大于第二预设匹配度(如可以第一预设匹配度为90％，第二预设匹配度为60％)，以便将识别匹配度分别与第一预设匹配度和第二预设匹配度进行比较，以确定其身份验证通过时的目标安全等级。
相应地，步骤S30中的基于待识别特征和标准识别特征获取对应的识别匹配度，具体包括如下步骤：
基于待识别文字特征和标准文字特征获取的第一识别匹配度；和/或，基于待识别语音特征和标准语音特征或者基于待识别语音特征和标准文字特征获取的第二识别匹配度。
识别匹配度是基于待识别特征和标准识别特征进行计算获取的用于体现两者相似度的指标。由于待识别特征包括待识别文字特征和/或待识别语音特征，标准识别特征包括标准文字特征和/或标准语音特征，可以根据待识别文字特征和标准文字特征获取第一识别匹配度；再根据待识别语音特征和标准语音特征获取第二识别匹配度，或者根据待识别语音特征和标准文字特征获取第二识别匹配度。
进一步地，基于待识别语音特征和标准语音特征或者基于待识别语音特征和标准文字特征获取的第二识别匹配度，具体包括如下两种情况：
第一种情况，采用预先训练好的声纹识别模型分别对待识别语音特征和标准语音特征进行声纹提取，分别获取待识别声纹特征和标准声纹特征；采用余弦相似度算法对待识别声纹特征和标准声纹特征进行余弦相似度计算，以获取第二识别匹配度。
其中，声纹识别模型预先训练并存储在服务器中的基于声纹进行说话人身份识别的模型。该声纹识别模型可以采用gmm-ubm(Gaussian mixture model-universal background model，即高斯混合模型-通用背景模型)，也可以采用其他模型训练得到。本实施例中，采用预先训练好的声纹识别模型分别对待识别语音特征和标准语音特征进行声纹提取，分别获取到待识别声纹特征和标准声纹特征，以便基于该待识别声纹特征和标准声纹特征进行身份识别。
进一步地，也可以预先采用该声纹识别模型对标准语音特征进行声纹提取，并将获取到的标准声纹特征存储在数据库中，在需要进行身份验证时，直接调取相应的标准声纹特征，以节省识别过程进行声纹特征提取的时间。
余弦相似度是用向量空间中两个向量夹角的余弦值作为衡量两个个体间差异大小的度量。余弦相似度算法是计算向量空间中两个向量夹角的余弦值的算法，可以用
Figure PCTCN2018081509-appb-000001
其中，a和b为向量空间中的待识别声纹特征和 标准声纹特征这两个向量，x和y是向量的坐标值。
本实施例中，采用预先训练好的声纹识别模型对待识别语音特征和标准语音特征进行声纹提取，以形成空间向量形式的待识别声纹特征和标准声纹特征，然后再采用余弦相似度算法对该待识别声纹特征和标准声纹特征进行余弦相似度计算，将获取的余弦值作为第二识别匹配度，使得获取到的第二识别匹配度可以很好地体现待识别语音特征和标准语音特征是否为同一客户的语音特征，以便进行身份验证。
第二种情况，采用预设的语音识别系统对待识别语音特征进行文字转换，获取待处理文字特征；采用余弦相似度算法对待处理文字特征和标准文字特征进行余弦相似度计算，以获取第二识别匹配度。
其中，语音识别系统是采用内置的语音识别算法把语音转变成文字的系统。服务器在获取智能手机上传的待识别语音特征之后，可以调用预设的语音识别系统对该待识别语音特征进行文字转换，获取待处理文字特征。本实施例中，预设的语音识别系统可以采用NLP(Natural Language Processing，自然语言处理)库，以将待识别语音特征转变成待处理文字特征。其中，NLP是人工智能的一个自然语言处理数据库，是用于将自然语言解析匹配为计算机能识别的机器语言的数据库。具体地，该NLP库可以采用斯坦福自然语音处理团队开发的Stanford CoreNLP，可以快速对待识别语音特征进行解析处理，以获取待处理文字特征。
具体地，采用余弦相似度算法对待处理文字特征和标准文字特征进行余弦相似度计算包括如下步骤：(1)对待处理文字特征和标准文字特征进行分词处理，如待处理文字特征分词后包括ABCDECF这七个词，标准文字特征分词后包括ABCGHEIF这八个词。(2)列出待处理文字特征和标准文字特征中出现的所有词，即包括ABCDEFGHI这九个词。(3)分别计算待处理文字特征和标准文字特征中出现的词的词频，则待处理文字特征的词频为A1，B1，C2，D1，E1，F1，G0，H0，I0；标准文字特征的词频为A1，B1，C1，D0，E1，F1，H1，G1，I1。(4)写出待处理文字特征和标准文字特征的词频向量，则待处理文字特征的词频向量为(1，1，2，1，1，1，0，0，0)，标准文字特征和词频向量为(1，1，1，0，1，1，1，1，1)。然后，采用上述的余弦相似度算法进行余弦相似度计算，计算过程如下：
Figure PCTCN2018081509-appb-000002
本实施例中，采用预设的语音识别系统对待识别语音特征进行文字转换，以获取待处理文字特征；然后，对待处理文字特征和标准文字特征分别进行分词、汇总词、词频计算并转换成词频向量，以便采用余弦相似度算法对两者转换成的词频向量进行余弦相似度计算，将获取的余弦值作为第二识别匹配度，使得获取的第二识别匹配度可以较清晰地反映客户基于安全保密问题回复的待识别语音特征是否与其预先存储的标准文字特征的匹配度，以便进行身份验证。
相应地，若识别匹配度大于或等于预设匹配度，则进入信息修改界面，具体包括如下几种情形：
第一种情形，若第一识别匹配度和第二识别匹配度均大于或等于第一预设匹配度，则进入高安全等级对应的信息修改界面。
如上所述，第一预设匹配度大于第二预设匹配度(如可以第一预设匹配度为90％，第二预设匹配度为60％)，若基于待识别特征和标准识别特征获取的识别匹配度大于第一预 设匹配度时，说明其确定的目标安全等级较高，为高安全等级。由于识别匹配度包括第一识别匹配度和第二识别匹配度，则若第一识别匹配度≥第一预设匹配度，且第二识别匹配度≥第一预设匹配度，说明其身份验证通过时的目标安全等级为高安全等级，此时可进入高安全等级对应的信息修改界面，在该信息修改界面上可显示数量较多的可修改主题及其对应的原始信息内容。
第二种情形，若第一识别匹配度和第二识别匹配度均小于第一预设匹配度且均大于或等于第二预设匹配度，则进入低安全等级对应的信息修改界面。
如上所述，第一预设匹配度大于第二预设匹配度(如可以第一预设匹配度为90％，第二预设匹配度为60％)，若基于待识别特征和标准识别特征获取的识别匹配度在第一预设匹配度和第二预设匹配度之间时，说明其虽然通过身份验证，但其目标安全等级较低，为低安全等级。由于识别匹配度包括第一识别匹配度和第二识别匹配度，则第一预设匹配度>第一识别匹配度≥第二预设匹配度，且第一预设匹配度>第二识别匹配度≥第二预设匹配度时，说明其身份验证通过时的目标安全等级为低安全等级，此时可进入低安全等级对应的信息修改界面，在该信息修改界面上可显示数量较少的可修改主题及其对应的原始信息内容。
第三种情形，若第一识别匹配度和第二识别匹配度中的任一个大于或等于第一预设匹配度，另一个小于第一预设匹配度且大于或等于第二预设匹配度，则进入中安全等级对应的信息修改界面。
中安全等级是界于高安全等级与低安全等级之间的安全等级。相应地，中安全等级对应的信息修改界面上显示的可修改主题的数量也是界于高安全等级和低安全等级对应的信息修改界面上显示的可修改主题的数量之间。如上所述，第一预设匹配度大于第二预设匹配度(如可以第一预设匹配度为90％，第二预设匹配度为60％)，识别匹配度包括第一识别匹配度和第二识别匹配度，在如下两种情况下可认定其身份验证通过时的目标安全等级为中安全等级：其一是，若第一预设匹配度>第一识别匹配度≥第二预设匹配度，且第二识别匹配度≥第一预设匹配度；其二是，第一预设匹配度>第二识别匹配度≥第二预设匹配度，且第一识别匹配度≥第一预设匹配度。
第四种情形，若第一识别匹配度和第二识别匹配度中的至少一种小于第二预设匹配度，则身份验证不通过，无法进入信息修改界面。
识别匹配度是基于待识别特征和标准识别特征进行计算获取的用于体现两者相似度的指标，第二预设匹配度是预先设置的数值较小的一个识别匹配度，是用于验证身份的最小指标，因此，第一识别匹配度和第二识别匹配度均需大于或等于第二预设匹配度。具体地，若第一识别匹配度小于第二预设匹配度，或者，第二识别匹配度小于第二预设匹配度，或者第一识别匹配度和第二识别匹配度同时小于第二预设匹配度，均会导致身份验证不通过，此时无法进入信息修改界面，以保障客户信息的安全。
该具体实施方式中，先基于待识别文字特征和标准文字特征获取第一识别匹配度，并基于待识别语音特征和标准语音特征或者基于待识别语音特征和标准文字特征获取的第二识别匹配度，以便基于第一识别匹配度和第二识别匹配度确定其身份验证通过时的目标安全等级。然后，将第一识别匹配度和第二识别匹配度分别与第一预设匹配度和第二预设匹配度进行比较，进而确定待识别特征和标准识别特征对应的目标安全等级为高安全等级、中安全等级和低安全等级中的任一个，分别进入与目标安全等级相对应的信息修改界面，使得客户信息的修改与其身份验证的安全等级相关联，以保障客户信息修改的安全性。
在一具体实施方式中，步骤S40中的获取目标修改数据，如图2所示，具体包括如下步骤：
S411：获取连续两次上传的语音修改数据。
语音修改数据是客户通过智能手机，以语音形式向服务器上传的用于体现所要修改的客户信息的数据，即客户上传的、以语音形式体现的所要修改的数据。为了避免因误触发而上传语音修改数据，或者为了避免上传错误的语音修改数据，需获取连续两次上传的语音修改数据，若连续两次上传的语音修改数据为相同的数据，则可以说明其为依据客户主观意愿上传的语音修改数据。本实施例中，可以预先设置起始标识符(如#*#)和结束标识符(如###)，并通过语音提示以指引客户进行相应操作，从而采集起始标识符和结束标识符之间的语音修改数据。
进一步地，可以限定连续两次上传的语音修改数据的时间间隔在预设时间内，以避免第一次上传语音修改数据之后间隔较长时间再上传另一次语音修改数据，从而导致系统处理进程或线程一直处于等待状态，影响服务器的正常工作。该预设时间为预先设置的时间区间，用于确定连续两次上传的语音修改数据是否为有效的数据。即，若两次上传的语音修改数据的时间间隔太久，大于预设时间时，则不认为两者为本实施例中所说的连续两次上传。
S412：对语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词。
在获取语音修改数据之后，调用预设的语音识别系统(如上提及的NLP)对语音修改数据进行文字转换，以将语音形式的语音修改数据转换成文字形式的原始修改数据。然后，采用预设的关键词提取算法(包括但不限于TextRank算法和TF-IDF算法等)对文字形式的原始修改数据进行关键词提取，以获取修改主题关键词和修改内容关键词。例如，客户上传的语音修改数据转换成原始修改数据为：我想将手机号修改为XXX，则其提取出的修改主题关键词为手机号，修改内容关键词为XXX。
S413：若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于修改主题关键词获取对应的目标修改主题。
第一可修改阈值是预先设置的与修改主题相关联的阈值，是用于评价能否确定修改主题的阈值。具体地，采用余弦相似度算法对连续两次上传的语音修改数据中的修改主题关键词进行余弦相似度计算，以获取两者的相似度，其计算过程如上对待处理文字特征和标准文字特征进行余弦相似度计算的过程一致，为避免赘述，不再一一详述计算过程。
本实施例中，可设第一可修改阈值为70％，若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于修改主题关键词获取对应的目标修改主题。例如手机号和手机号码这两个连续上传的修改主题关键词的相似度为75％，大于第一可修改阈值70％，此时，可基于手机号或手机号码查找数据库中，以确定与其对应的客户信息中的可修改主题，作为目标修改主题。又如，手机号和证件号这两个连续上传的修改主题关键词的相似度为33.3％，小于第一可修改阈值，则不能基于修改主题关键词获取对应的目标修改主题。
S414：若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于修改内容关键词获取对应的目标修改内容。
第二可修改阈值是预先设置的与修改内容相关联的阈值，是用于评价能否确定修改内容的阈值。该第二可修改阈值可与第一可修改阈值一致，也可以不一致，根据所要修改的 内容进行确定。具体地，采用余弦相似度算法对连续两次上传的语音修改数据中的修改内容关键词进行余弦相似度计算，以获取两者的相似度，为避免赘述，不再一一详述计算过程。由于修改内容涉及客户信息的实质内容，一般要求其第二可修改阈值设置较大，甚至接近100％，尤其是涉及手机号、证件号、车架号、发动机号和车牌等值时。
S415：基于目标修改主题和目标修改内容，获取目标修改数据。
可以理解地，只有根据步骤S413获取目标修改主题和根据步骤S414获取目标修改内容之后，才可基于目标修改主题和目标修改内容，获取目标修改数据。即若连续两次上传的语音修改数据中的修改主题关键词的相似度未达到第一可修改阈值，或者连续两次上传的语音修改数据中的修改内容关键词的相似度未达到第二可修改阈值，均导致没有获取目标修改数据，需重新执行步骤S411-S415，或者采用其他方式确定目标修改数据。
该具体实施方式中，通过获取连续两次上传的语音修改数据，以避免误触发上传或者上传错误，从而保证客户信息的修改符合客户的真实意愿，保证客户信息的安全。通过对语音修改数据进行文字转换和关键词提取，以获取修改主题关键词和修改内容关键词；再基于连续两次获取到的修改主题关键词计算其相似度，并将该相似度与第一可修改阈值进行比较，以确定目标修改主题；再基于连续两次获取到的修改内容关键词计算其相似度，并将该相似度与第二可修改阈值进行比较，以获取目标修改内容。在目标修改主题和目标修改内容同时确定时，才可获取到相应的目标修改数据，以保证目标修改数据的真实性，确保客户信息修改的安全。
在一具体实施方式中，若客户连续上传多次(大于两次)语音修改数据，在重复执行步骤S411-S415之后，仍无法同时确定其目标修改主题和目标修改内容，即每一组连续两次上传的语音修改数据进行处理后，其语音修改数据中的修改主题关键词的相似度未达到第一可修改阈值，或者其语音修改数据中的修改内容关键词的相似度未达到第二可修改阈值，需先执行步骤S411-S415进行重复验证，在重复验证的次数达到预设次数(预先自主设置的次数，可以为2次、3次或其他次数)时，可以由坐席人员进行误差修改，以便于确定目标修改数据。即步骤S40中的获取目标修改数据，如图3所示，具体包括如下步骤：
S421：获取误差修改数据，误差修改数据包括误差修改主题和对应的误差修改内容。
误差修改数据是坐席人员通过坐席终端向服务器发送的用于体现要修改的客户信息的数据。误差修改主题是坐席人员在与客户通话中，根据客户多次上传的语音修改数据确定其所要修改的客户信息中的主题。误差修改内容是与误差修改主题相对应的具体值或具体内容。
S422：基于误差修改主题获取对应的修改限制规则和原始修改内容。
修改限制规则与误差修改主题相对应的用于限制坐席人员修改数据的规则。修改限制规则可以依据修改位数、类似拼音和模糊音等进行限制。基于修改位数的修改限制规则可以设置如下：手机号修改时，其修改位数控制在2位；或者车架号修改时，其修改位数控制在5位等。其中，修改位数可由误差修改主题对应的值的位数自主设置。基于类似拼音的修改限制规则可以用于实现拼音相同的文字相互替换。基于模糊音的修改限制规则可以用于实现前后鼻音和翘舌音等模糊音的等位替换。原始修改内容是客户采用步骤S411-S415确定的与误差修改主题相对应的修改内容，原始修改内容可能是一个，也可能是多个，即客户在多次上传的语音修改数据中识别出的修改内容，但这些修改内容没有在连续两次上传时相同。
S423：采用修改限制规则对原始修改内容和误差修改内容进行检验处理，若检验通过， 则将误差修改内容作为目标修改内容。
对每一误差修改主题对应的误差修改内容和原始修改内容，采用步骤S422确定的修改限制规则进行检验，并将检验通过时的误差修改内容作为目标修改内容。如误差修改主题对应的修改限制规则为基于位数的修改限制规则时，判断误差修改内容和原始修改内容的位数差是否在预先设置的修改位数内，若是，则检验通过，并将误差修改内容作为目标修改内容。
步骤S421-S423中，可以使客户在与坐席人员通话中修改客户信息的过程中，因为发音或其他原因导致采用步骤S411-SS415无法验证通过时，可以由坐席人员进行协助修改，以帮助客户进行客户信息修改；并通过设置修改限制规则，用于限制坐席人员修改客户信息，以起到防止恶意修改的目的。
应理解，上述实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
实施例2
图4示出与实施例1中通话中客户信息修改方法一一对应的通话中客户信息修改装置的原理框图。如图4所示，该通话中客户信息修改装置包括修改验证请求获取模块10、标准识别特征获取模块20、识别匹配处理模块30、修改数据获取显示模块40和修改确认处理模块50。其中，修改验证请求获取模块10、标准识别特征获取模块20、识别匹配处理模块30、修改数据获取显示模块40和修改确认处理模块50的实现功能与实施例1中通话中客户信息修改方法对应的步骤一一对应，为避免赘述，本实施例不一一详述。
修改验证请求获取模块10，用于获取修改验证请求，修改验证请求包括客户手机号和待识别特征。
标准识别特征获取模块20，用于基于客户手机号查询数据库，获取对应的标准识别特征。
识别匹配处理模块30，用于基于待识别特征和标准识别特征获取对应的识别匹配度，若识别匹配度大于或等于预设匹配度，则进入信息修改界面。
修改数据获取显示模块40，用于获取目标修改数据，在信息修改界面中显示目标修改数据，设置目标修改数据为不可修改状态。
修改确认处理模块50，用于获取修改确认指令，将目标修改数据上传到数据库。
优选地，目标修改数据包括目标修改主题和对应的目标修改内容。
识别匹配处理模块30，用于若识别匹配度大于或等于预设匹配度，则基于识别匹配度和预设匹配度获取对应的目标安全等级，并进入与目标安全等级相对应的信息修改界面，信息修改界面上显示与目标安全等级相对应的可修改主题和对应的原始信息内容。
修改数据获取显示模块40，用于若目标修改主题属于可修改主题中的主题时，则在信息修改界面上对比显示与目标修改主题相对应的原始信息内容和目标修改内容。
优选地，待识别特征包括待识别文字特征和/或待识别语音特征。
标准识别特征包括标准文字特征和/或标准语音特征。
预设匹配度包括第一预设匹配度和第二预设匹配度，第一预设匹配度大于第二预设匹配度。
目标安全等级具体为高安全等级、中安全等级和低安全等级中的任一个。
识别匹配处理模块30包括识别匹配度获取单元311，用于基于待识别文字特征和标准文字特征获取的第一识别匹配度；和/或，基于待识别语音特征和标准语音特征或者基 于待识别语音特征和标准文字特征获取的第二识别匹配度。
识别匹配处理模块30还包括第一匹配处理单元321、第二匹配处理单元322和第三匹配处理单元323。
第一匹配处理单元321，用于若第一识别匹配度和第二识别匹配度均大于或等于第一预设匹配度，则进入高安全等级对应的信息修改界面。
第二匹配处理单元322，用于若第一识别匹配度和第二识别匹配度均小于第一预设匹配度且均大于或等于第二预设匹配度，则进入低安全等级对应的信息修改界面。
第三匹配处理单元323，用于若第一识别匹配度和第二识别匹配度中的任一个大于或等于第一预设匹配度，另一个小于第一预设匹配度且大于或等于第二预设匹配度，则进入中安全等级对应的信息修改界面。
优选地，识别匹配度获取单元311包括声纹匹配处理子单元3111或者语音匹配处理子单元3112。
声纹匹配处理子单元3111，用于采用预先训练好的声纹识别模型分别对待识别语音特征和标准语音特征进行声纹提取，分别获取待识别声纹特征和标准声纹特征；采用余弦相似度算法对待识别声纹特征和标准声纹特征进行余弦相似度计算，以获取第二识别匹配度。或者，
语音匹配处理子单元3112，用于采用预设的语音识别系统对待识别语音特征进行文字转换，获取待处理文字特征；采用余弦相似度算法对待处理文字特征和标准文字特征进行余弦相似度计算，以获取第二识别匹配度。
优选地，修改数据获取显示模块40包括语音修改数据获取单元411、关键词提取获取单元412、目标修改主题获取单元413、目标修改内容获取单元414和目标修改数据获取单元415。
语音修改数据获取单元411，用于获取连续两次上传的语音修改数据。
关键词提取获取单元412，用于对语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词。
目标修改主题获取单元413，用于若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于修改主题关键词获取对应的目标修改主题。
目标修改内容获取单元414，用于若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于修改内容关键词获取对应的目标修改内容。
目标修改数据获取单元415，用于基于目标修改主题和目标修改内容，获取目标修改数据。
优选地，修改数据获取显示模块40包括误差修改数据获取单元421、限制规则和内容获取单元422和修改验证处理单元423。
误差修改数据获取单元421，用于获取误差修改数据，误差修改数据包括误差修改主题和对应的误差修改内容。
限制规则和内容获取单元422，用于基于误差修改主题获取对应的修改限制规则和原始修改内容。
修改验证处理单元423，用于采用修改限制规则对原始修改内容和误差修改内容进行检验处理，若检验通过，则将误差修改内容作为目标修改内容。
实施例3
本实施例提供一个或多个存储有计算机可读指令的非易失性可读存储介质。该一个或 多个存储有计算机可读指令的非易失性可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行时实现实施例1中通话中客户信息修改方法，为避免重复，这里不再赘述。或者，该计算机可读指令被处理器执行时实现实施例2中通话中客户信息修改中各模块/单元的功能，为避免重复，这里不再赘述。
可以理解地，一个或多个存储有计算机可读指令的非易失性可读存储介质可以包括：能够携带所述计算机可读指令的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、电载波信号和电信信号等。
实施例4
图5是本申请一实施例提供的计算机设备的示意图。如图5所示，该实施例的计算机设备60包括：处理器61、存储器62以及存储在存储器62中并可在处理器61上运行的计算机可读指令63。处理器61执行计算机可读指令63时实现上述实施例1中的通话中客户信息修改方法的步骤，例如图1所示的步骤S10-S50。或者，处理器61执行计算机可读指令63时实现上述实施例2中的通话中客户信息修改装置的模块/单元的功能，例如图4所示的修改验证请求获取模块10、标准识别特征获取模块20、识别匹配处理模块30、修改数据获取显示模块40和修改确认处理模块50的功能。
示例性的，计算机可读指令63可以被分割成一个或多个模块/单元，一个或者多个模块/单元被存储在存储器62中，并由处理器61执行，以完成本申请。一个或多个模块/单元可以是能够完成特定功能的一系列计算机可读指令的指令段段，该指令段用于描述计算机可读指令63在计算机设备60中的执行过程。例如，计算机可读指令63可以被分割成修改验证请求获取模块10、标准识别特征获取模块20、识别匹配处理模块30、修改数据获取显示模块40和修改确认处理模块50，各模块具体功能如实施例2中所述，在此不一一赘述。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。
以上所述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。

Claims (20)

  1. 一种通话中客户信息修改方法，其特征在于，包括：
    获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
    基于所述客户手机号查询数据库，获取对应的标准识别特征；
    基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
    获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
    获取修改确认指令，将所述目标修改数据上传到所述数据库。
  2. 如权利要求1所述的通话中客户信息修改方法，其特征在于，所述目标修改数据包括目标修改主题和对应的目标修改内容；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述识别匹配度大于或等于预设匹配度，则基于所述识别匹配度和所述预设匹配度获取对应的目标安全等级，并进入与所述目标安全等级相对应的信息修改界面，所述信息修改界面上显示与所述目标安全等级相对应的可修改主题和对应的原始信息内容；
    所述在所述信息修改界面中显示所述目标修改数据，包括：
    若所述目标修改主题属于所述可修改主题中的主题时，则在所述信息修改界面上对比显示与所述目标修改主题相对应的原始信息内容和目标修改内容。
  3. 如权利要求2所述的通话中客户信息修改方法，其特征在于，所述待识别特征包括待识别文字特征和/或待识别语音特征；
    所述标准识别特征包括标准文字特征和/或标准语音特征；
    所述预设匹配度包括第一预设匹配度和第二预设匹配度，所述第一预设匹配度大于所述第二预设匹配度；
    所述目标安全等级具体为高安全等级、中安全等级和低安全等级中的任一个；
    所述基于所述待识别特征和标准识别特征获取对应的识别匹配度，包括：
    基于所述待识别文字特征和所述标准文字特征获取的第一识别匹配度；和/或，基于所述待识别语音特征和所述标准语音特征或者基于所述待识别语音特征和所述标准文字特征获取的第二识别匹配度；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述第一识别匹配度和所述第二识别匹配度均大于或等于所述第一预设匹配度，则进入高安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度均小于所述第一预设匹配度且均大于或等于所述第二预设匹配度，则进入低安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度中的任一个大于或等于所述第一预设匹配度，另一个小于所述第一预设匹配度且大于或等于所述第二预设匹配度，则进入中安全等级对应的信息修改界面。
  4. 如权利要求3所述的通话中客户信息修改方法，其特征在于，所述基于所述待识别语音特征和所述标准语音特征或者基于所述所述待识别语音特征和标准文字特征获取的第二识别匹配度，包括：
    采用预先训练好的声纹识别模型分别对所述待识别语音特征和所述标准语音特征进 行声纹提取，分别获取待识别声纹特征和标准声纹特征；采用余弦相似度算法对所述待识别声纹特征和所述标准声纹特征进行余弦相似度计算，以获取所述第二识别匹配度；或者，
    采用预设的语音识别系统对所述待识别语音特征进行文字转换，获取待处理文字特征；采用余弦相似度算法对所述待处理文字特征和所述标准文字特征进行余弦相似度计算，以获取所述第二识别匹配度。
  5. 如权利要求1所述的通话中客户信息修改方法，其特征在于，所述获取目标修改数据，包括：
    获取连续两次上传的语音修改数据；
    对所述语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词；
    若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于所述修改主题关键词获取对应的目标修改主题；
    若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于所述修改内容关键词获取对应的目标修改内容；
    基于所述目标修改主题和所述目标修改内容，获取所述目标修改数据。
  6. 如权利要求5所述的通话中客户信息修改方法，其特征在于，所述获取目标修改数据，包括：
    获取误差修改数据，所述误差修改数据包括误差修改主题和对应的误差修改内容；
    基于所述误差修改主题获取对应的修改限制规则和原始修改内容；
    采用所述修改限制规则对所述原始修改内容和所述误差修改内容进行检验处理，若检验通过，则将所述误差修改内容作为所述目标修改内容。
  7. 一种通话中客户信息修改装置，其特征在于，包括：
    修改验证请求获取模块，用于获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
    标准识别特征获取模块，用于基于所述客户手机号查询数据库，获取对应的标准识别特征；
    识别匹配处理模块，用于基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
    修改数据获取显示模块，用于获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
    修改确认处理模块，用于获取修改确认指令，将所述目标修改数据上传到所述数据库。
  8. 如权利要求7所述的通话中客户信息修改装置，其特征在于，所述修改数据获取显示模块包括：
    语音修改数据获取单元，用于获取连续两次上传的语音修改数据；
    关键词提取获取单元，用于对所述语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词；
    目标修改主题获取单元，用于若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于所述修改主题关键词获取对应的目标修改主题；
    目标修改内容获取单元，用于若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于所述修改内容关键词获取对应的目标修改内容；
    目标修改数据获取单元，用于基于所述目标修改主题和所述目标修改内容，获取所述 目标修改数据。
  9. 一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，其特征在于，所述处理器执行所述计算机可读指令时实现如下步骤：
    获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
    基于所述客户手机号查询数据库，获取对应的标准识别特征；
    基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
    获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
    获取修改确认指令，将所述目标修改数据上传到所述数据库。
  10. 如权利要求9所述的计算机设备，其特征在于，所述目标修改数据包括目标修改主题和对应的目标修改内容；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述识别匹配度大于或等于预设匹配度，则基于所述识别匹配度和所述预设匹配度获取对应的目标安全等级，并进入与所述目标安全等级相对应的信息修改界面，所述信息修改界面上显示与所述目标安全等级相对应的可修改主题和对应的原始信息内容；
    所述在所述信息修改界面中显示所述目标修改数据，包括：
    若所述目标修改主题属于所述可修改主题中的主题时，则在所述信息修改界面上对比显示与所述目标修改主题相对应的原始信息内容和目标修改内容。
  11. 如权利要求10所述的计算机设备，其特征在于，所述待识别特征包括待识别文字特征和/或待识别语音特征；
    所述标准识别特征包括标准文字特征和/或标准语音特征；
    所述预设匹配度包括第一预设匹配度和第二预设匹配度，所述第一预设匹配度大于所述第二预设匹配度；
    所述目标安全等级具体为高安全等级、中安全等级和低安全等级中的任一个；
    所述基于所述待识别特征和标准识别特征获取对应的识别匹配度，包括：
    基于所述待识别文字特征和所述标准文字特征获取的第一识别匹配度；和/或，基于所述待识别语音特征和所述标准语音特征或者基于所述待识别语音特征和所述标准文字特征获取的第二识别匹配度；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述第一识别匹配度和所述第二识别匹配度均大于或等于所述第一预设匹配度，则进入高安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度均小于所述第一预设匹配度且均大于或等于所述第二预设匹配度，则进入低安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度中的任一个大于或等于所述第一预设匹配度，另一个小于所述第一预设匹配度且大于或等于所述第二预设匹配度，则进入中安全等级对应的信息修改界面。
  12. 如权利要求11所述的计算机设备，其特征在于，所述基于所述待识别语音特征和所述标准语音特征或者基于所述所述待识别语音特征和标准文字特征获取的第二识别匹配度，包括：
    采用预先训练好的声纹识别模型分别对所述待识别语音特征和所述标准语音特征进行声纹提取，分别获取待识别声纹特征和标准声纹特征；采用余弦相似度算法对所述待识别声纹特征和所述标准声纹特征进行余弦相似度计算，以获取所述第二识别匹配度；或者，
    采用预设的语音识别系统对所述待识别语音特征进行文字转换，获取待处理文字特征；采用余弦相似度算法对所述待处理文字特征和所述标准文字特征进行余弦相似度计算，以获取所述第二识别匹配度。
  13. 如权利要求9所述的计算机设备，其特征在于，所述获取目标修改数据，包括：
    获取连续两次上传的语音修改数据；
    对所述语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词；
    若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于所述修改主题关键词获取对应的目标修改主题；
    若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于所述修改内容关键词获取对应的目标修改内容；
    基于所述目标修改主题和所述目标修改内容，获取所述目标修改数据。
  14. 如权利要求13所述的计算机设备，其特征在于，所述获取目标修改数据，包括：
    获取误差修改数据，所述误差修改数据包括误差修改主题和对应的误差修改内容；
    基于所述误差修改主题获取对应的修改限制规则和原始修改内容；
    采用所述修改限制规则对所述原始修改内容和所述误差修改内容进行检验处理，若检验通过，则将所述误差修改内容作为所述目标修改内容。
  15. 一个或多个存储有计算机可读指令的非易失性可读存储介质，其特征在于，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤：
    获取修改验证请求，所述修改验证请求包括客户手机号和待识别特征；
    基于所述客户手机号查询数据库，获取对应的标准识别特征；
    基于所述待识别特征和标准识别特征获取对应的识别匹配度，若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面；
    获取目标修改数据，在所述信息修改界面中显示所述目标修改数据，设置所述目标修改数据为不可修改状态；
    获取修改确认指令，将所述目标修改数据上传到所述数据库。
  16. 如权利要求15所述的非易失性可读存储介质，其特征在于，所述目标修改数据包括目标修改主题和对应的目标修改内容；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述识别匹配度大于或等于预设匹配度，则基于所述识别匹配度和所述预设匹配度获取对应的目标安全等级，并进入与所述目标安全等级相对应的信息修改界面，所述信息修改界面上显示与所述目标安全等级相对应的可修改主题和对应的原始信息内容；
    所述在所述信息修改界面中显示所述目标修改数据，包括：
    若所述目标修改主题属于所述可修改主题中的主题时，则在所述信息修改界面上对比显示与所述目标修改主题相对应的原始信息内容和目标修改内容。
  17. 如权利要求16所述的非易失性可读存储介质，其特征在于，所述待识别特征包括待识别文字特征和/或待识别语音特征；
    所述标准识别特征包括标准文字特征和/或标准语音特征；
    所述预设匹配度包括第一预设匹配度和第二预设匹配度，所述第一预设匹配度大于所述第二预设匹配度；
    所述目标安全等级具体为高安全等级、中安全等级和低安全等级中的任一个；
    所述基于所述待识别特征和标准识别特征获取对应的识别匹配度，包括：
    基于所述待识别文字特征和所述标准文字特征获取的第一识别匹配度；和/或，基于所述待识别语音特征和所述标准语音特征或者基于所述待识别语音特征和所述标准文字特征获取的第二识别匹配度；
    所述若所述识别匹配度大于或等于预设匹配度，则进入信息修改界面，包括：
    若所述第一识别匹配度和所述第二识别匹配度均大于或等于所述第一预设匹配度，则进入高安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度均小于所述第一预设匹配度且均大于或等于所述第二预设匹配度，则进入低安全等级对应的信息修改界面；
    若所述第一识别匹配度和所述第二识别匹配度中的任一个大于或等于所述第一预设匹配度，另一个小于所述第一预设匹配度且大于或等于所述第二预设匹配度，则进入中安全等级对应的信息修改界面。
  18. 如权利要求17所述的非易失性可读存储介质，其特征在于，所述基于所述待识别语音特征和所述标准语音特征或者基于所述所述待识别语音特征和标准文字特征获取的第二识别匹配度，包括：
    采用预先训练好的声纹识别模型分别对所述待识别语音特征和所述标准语音特征进行声纹提取，分别获取待识别声纹特征和标准声纹特征；采用余弦相似度算法对所述待识别声纹特征和所述标准声纹特征进行余弦相似度计算，以获取所述第二识别匹配度；或者，
    采用预设的语音识别系统对所述待识别语音特征进行文字转换，获取待处理文字特征；采用余弦相似度算法对所述待处理文字特征和所述标准文字特征进行余弦相似度计算，以获取所述第二识别匹配度。
  19. 如权利要求15所述的非易失性可读存储介质，其特征在于，所述获取目标修改数据，包括：
    获取连续两次上传的语音修改数据；
    对所述语音修改数据进行文字转换和关键词提取，获取修改主题关键词和修改内容关键词；
    若连续两次上传的语音修改数据中的修改主题关键词的相似度达到第一可修改阈值，则基于所述修改主题关键词获取对应的目标修改主题；
    若连续两次上传的语音修改数据中的修改内容关键词的相似度达到第二可修改阈值，则基于所述修改内容关键词获取对应的目标修改内容；
    基于所述目标修改主题和所述目标修改内容，获取所述目标修改数据。
  20. 如权利要求19所述的非易失性可读存储介质，其特征在于，所述获取目标修改数据，包括：
    获取误差修改数据，所述误差修改数据包括误差修改主题和对应的误差修改内容；
    基于所述误差修改主题获取对应的修改限制规则和原始修改内容；
    采用所述修改限制规则对所述原始修改内容和所述误差修改内容进行检验处理，若检验通过，则将所述误差修改内容作为所述目标修改内容。
PCT/CN2018/081509 2018-03-12 2018-04-02 通话中客户信息修改方法、装置、计算机设备及存储介质 WO2019174073A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810198383.6 2018-03-12
CN201810198383.6A CN108510290B (zh) 2018-03-12 2018-03-12 通话中客户信息修改方法、装置、计算机设备及存储介质

Publications (1)

Publication Number Publication Date
WO2019174073A1 true WO2019174073A1 (zh) 2019-09-19

Family

ID=63377248

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/081509 WO2019174073A1 (zh) 2018-03-12 2018-04-02 通话中客户信息修改方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN108510290B (zh)
WO (1) WO2019174073A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112749956A (zh) * 2020-07-20 2021-05-04 腾讯科技（深圳）有限公司 信息处理方法、装置及设备

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110706696A (zh) * 2019-09-25 2020-01-17 珠海格力电器股份有限公司 一种语音控制方法及装置
CN111210826B (zh) * 2019-12-26 2022-08-05 深圳市优必选科技股份有限公司 语音信息处理方法、装置、存储介质和智能终端
CN111554296B (zh) * 2020-04-27 2023-11-10 中国银行股份有限公司 一种客户信息修改方法、装置、服务器及存储介质
CN112201245B (zh) * 2020-09-30 2024-02-06 中国银行股份有限公司 信息处理方法、装置、设备及存储介质
CN112491840B (zh) * 2020-11-17 2023-07-07 平安养老保险股份有限公司 信息修改方法、装置、计算机设备及存储介质
CN114489897B (zh) * 2022-01-21 2023-08-08 北京字跳网络技术有限公司 一种对象处理方法、装置、终端设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106027276A (zh) * 2016-01-21 2016-10-12 李明 一种内管服务器
US20160323274A1 (en) * 2015-04-30 2016-11-03 Google Inc. Facial Profile Password to Modify User Account Data for Hands-Free Transactions
WO2016176832A1 (zh) * 2015-05-06 2016-11-10 华为技术有限公司 鉴权方法和接入设备
CN107196971A (zh) * 2017-07-19 2017-09-22 中国银行股份有限公司 信息处理方法、装置、电子设备及服务器
CN107645467A (zh) * 2016-07-20 2018-01-30 平安科技（深圳）有限公司 密码修改方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160323274A1 (en) * 2015-04-30 2016-11-03 Google Inc. Facial Profile Password to Modify User Account Data for Hands-Free Transactions
WO2016176832A1 (zh) * 2015-05-06 2016-11-10 华为技术有限公司 鉴权方法和接入设备
CN106027276A (zh) * 2016-01-21 2016-10-12 李明 一种内管服务器
CN107645467A (zh) * 2016-07-20 2018-01-30 平安科技（深圳）有限公司 密码修改方法和装置
CN107196971A (zh) * 2017-07-19 2017-09-22 中国银行股份有限公司 信息处理方法、装置、电子设备及服务器

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112749956A (zh) * 2020-07-20 2021-05-04 腾讯科技（深圳）有限公司 信息处理方法、装置及设备

Also Published As

Publication number Publication date
CN108510290B (zh) 2019-08-06
CN108510290A (zh) 2018-09-07

Similar Documents

Publication Publication Date Title
WO2019174073A1 (zh) 通话中客户信息修改方法、装置、计算机设备及存储介质
US10853676B1 (en) Validating identity and/or location from video and/or audio
US10650379B2 (en) Method and system for validating personalized account identifiers using biometric authentication and self-learning algorithms
WO2020077896A1 (zh) 提问数据生成方法、装置、计算机设备和存储介质
WO2020024398A1 (zh) 生物特征辅助支付方法、装置、计算机设备及存储介质
WO2019153982A1 (zh) 支付处理方法、装置、计算机设备和存储介质
WO2020077885A1 (zh) 身份验证方法、装置、计算机设备和存储介质
US20190294900A1 (en) Remote user identity validation with threshold-based matching
EP3042333A1 (en) Biometric verification using predicted signatures
CN109033150B (zh) 敏感词验证方法、装置、计算机设备及存储介质
WO2020034761A1 (zh) 核身产品推送及核身方法和系统
US20200374286A1 (en) Real time selfie systems and methods for automating user identify verification
WO2019019627A1 (zh) 电子保单的状态查询方法、装置、用户终端及存储介质
WO2019227576A1 (zh) 发票校验方法、装置、计算机设备及存储介质
US20160012823A1 (en) System and methods for personal identification number authentication and verification
CN109685514A (zh) 用于支付的方法、装置、存储介质及电子设备
CN110324314A (zh) 用户注册方法及装置、存储介质、电子设备
CN107147499A (zh) 使用语音输入进行验证的方法和系统
US20220335433A1 (en) Biometrics-Infused Dynamic Knowledge-Based Authentication Tool
CN114493902A (zh) 多模态信息异常监控方法、装置、计算机设备及存储介质
US20240013214A1 (en) Method for Determining the Likelihood for Someone to Remember a Particular Transaction
CN112201254A (zh) 无感语音认证方法、装置、设备及存储介质
WO2019242401A1 (zh) 用户操作权限的控制方法、装置、存储介质及电子设备
CN111916086B (zh) 语音交互控制方法、装置、计算机设备和存储介质
JP2023530893A (ja) データ処理および取引決定システム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18909372

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 09.12.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18909372

Country of ref document: EP

Kind code of ref document: A1