WO2022105496A1 - 智能回访方法、装置、电子设备及可读存储介质 - Google Patents

智能回访方法、装置、电子设备及可读存储介质 Download PDF

Info

Publication number
WO2022105496A1
WO2022105496A1 PCT/CN2021/123905 CN2021123905W WO2022105496A1 WO 2022105496 A1 WO2022105496 A1 WO 2022105496A1 CN 2021123905 W CN2021123905 W CN 2021123905W WO 2022105496 A1 WO2022105496 A1 WO 2022105496A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
return visit
returned
target
feature
Prior art date
Application number
PCT/CN2021/123905
Other languages
English (en)
French (fr)
Inventor
郭锦宏
Original Assignee
深圳壹账通智能科技有限公司
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 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2022105496A1 publication Critical patent/WO2022105496A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/23Clustering techniques
    • G06F18/232Non-hierarchical techniques
    • G06F18/2321Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/30Authentication, i.e. establishing the identity or authorisation of security principals
    • G06F21/31User authentication
    • G06F21/32User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201Market modelling; Market analysis; Collecting market data
    • G06Q30/0203Market surveys; Market polls

Definitions

  • the present application relates to the field of artificial intelligence technology and data analysis, and in particular, to an intelligent return visit method, device, electronic device and readable storage medium.
  • the intelligent return visit method provided in this application includes:
  • the characteristic factor corresponding to the return visit category obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • the identity verification is performed on the user to be returned based on the target return visit channel, and when the identity verification is passed, the return visit is performed on the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • the present application also provides an intelligent return visit device, the device includes:
  • a parsing module configured to parse the return visit request sent by the first client, obtain the identifier of the user to be returned carried in the return visit request, and obtain user data of the user to be returned from the first database based on the identifier;
  • a generating module is used to determine the return visit category corresponding to the user to be returned based on the user data, and to establish a user portrait for the user to be returned, and to generate a return visit for the user to be returned based on the user portrait and the return visit category. questionnaire;
  • a determination module configured to acquire the characteristic factor corresponding to the return visit category, obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • a return visit module configured to perform identity verification on the user to be returned based on the target return visit channel, and when the identity verification is passed, return visit to the user to be returned based on the target return visit channel and the return visit volume, and generate a return visit Report.
  • the present application also provides an electronic device, the electronic device comprising:
  • the memory stores an intelligent return access program executable by the at least one processor, and the intelligent return access program is executed by the at least one processor, so that the at least one processor can perform the following steps:
  • the characteristic factor corresponding to the return visit category obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • the identity verification is performed on the user to be returned based on the target return visit channel, and when the identity verification is passed, the return visit is performed on the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • the present application also provides a computer-readable storage medium, where an intelligent return visit program is stored on the computer-readable storage medium, and the intelligent return visit program can be executed by one or more processors to implement the following steps:
  • the characteristic factor corresponding to the return visit category obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • the identity verification is performed on the user to be returned based on the target return visit channel, and when the identity verification is passed, the return visit is performed on the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • FIG. 1 is a schematic flowchart of an intelligent return visit method provided by an embodiment of the present application.
  • FIG. 2 is a schematic diagram of a module of an intelligent return visit device provided by an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram of an electronic device for implementing an intelligent return visit method provided by an embodiment of the present application
  • the embodiments of the present application may acquire and process related data based on artificial intelligence technology.
  • Artificial Intelligence is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .
  • the basic technologies of artificial intelligence generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics.
  • Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning/deep learning.
  • the present application provides an intelligent return visit method.
  • FIG. 1 it is a schematic flowchart of an intelligent return visit method according to an embodiment of the present application.
  • the method may be performed by an electronic device, which may be implemented by software and/or hardware.
  • the intelligent return visit method includes:
  • the first client may be the backend server of the insurance company, or may be a client (work computer or mobile phone) corresponding to an employee of the insurance company.
  • the identifier of the user to be returned may be the ID number of the user, and the identifier of the user may be the identifier of one user or the identifiers of a group of users.
  • the user data includes the user's basic information data and the user's product information data.
  • the basic information includes name, age, address, occupation, income, etc.
  • the product information data includes the insurance type purchased by the user, insurance premium, and insurance payment. Frequency, sales channels, time of purchase, etc.
  • the determining the return visit category corresponding to the user to be returned based on the user data includes: determining the return visit category corresponding to the user to be returned according to the insurance category in the product information data.
  • the return visit category corresponding to user 1 is critical illness insurance return visit
  • the return visit category corresponding to user 2 is participating insurance return visit.
  • the preset keyword set is matched with the user data to obtain index data of multiple dimensions of the user, so as to establish the user portrait of the user to be returned to visit.
  • the preset keyword set may be age, income, insurance type, payment frequency
  • the user portrait corresponding to user 1 may be ⁇ 50 years old, 10,000, health insurance, payment by year ⁇
  • the user corresponding to user 2 The portrait can be ⁇ 25 years old, 8000, participating insurance, paid by year ⁇ .
  • the generating a return visit volume for the user to be returned based on the user portrait and the return visit category includes:
  • A1 extracting data corresponding to preset index items from the user data to generate a first questionnaire
  • the preset indicator items may be the user's mobile phone number and communication address. It is assumed that the mobile phone number recorded in the user data of user 1 is 135xxxxxxxx, and the communication address is No. , then the generated first questionnaire of User 1 can be "Is your mobile phone number 135xxxxxxxx? If not, please provide your new mobile phone number", "Excuse me, your mailing address is xx hometown x, xx district, Shenzhen, Guangdong province. Is the building xx number? If not, please provide a new mailing address.” Through the first questionnaire, it can be judged whether the communication information of the user to be returned has been updated.
  • A2 Obtain the second questionnaire corresponding to the return visit category from the second database, and extract a third questionnaire from the second questionnaire based on the user portrait;
  • the standard questionnaire corresponding to each type of return visit is pre-stored in the second database.
  • the content of the standard questionnaire is relatively comprehensive and includes many return visit questions.
  • the return visit questions in the second questionnaire need to be selected and processed in a targeted manner. to improve return visit efficiency.
  • user 1 may pay more attention to the compensation period and compensation ratio, so the questions related to the compensation period and compensation ratio are selected from the second questionnaire corresponding to user 1 as the third questionnaire of user 1; user 2 It is possible to pay more attention to the exemption clause and historical interest rate, then the questions related to the exemption clause and the historical interest rate are selected from the second questionnaire corresponding to user 2 as the third questionnaire of user 2.
  • the obtained return interview questionnaire is more relevant to the user, which reduces the amount of return interview questions while meeting the user's needs, and can speed up the return interview efficiency.
  • This embodiment pre-sets corresponding characteristic factors for each type of return visit.
  • the characteristic factors corresponding to participating insurance may be premiums, income, historical premiums, and age.
  • the characteristic values corresponding to the above characteristic factors can be determined based on user data (for example, , if the premium is 5,000 yuan, the eigenvalue corresponding to the premium may be 5,000.
  • the value after discretization and normalization of 5,000 can also be used as the eigenfactor corresponding to the premium, and the discretization and normalization
  • an array formed by eigenvalues corresponding to each eigenfactor is used as the first feature.
  • the determining of the target return visit channel corresponding to the user to be returned based on the first feature includes:
  • the PCA Principal Component Analysis
  • the PCA algorithm is used for the original feature. For all variables, delete redundant variables from the repeated variables (closely related variables), and establish as few new variables as possible, so that the new variables are not related to each other, and the new variables retain the original information as much as possible, that is, the m-dimensional feature map To n dimensions (n ⁇ m), this step can make subsequent feature processing more efficient.
  • the third feature is calculated based on the feature factor and user data of each user in the third database, and the calculation process is the same as the calculation process of the first feature, which is not repeated here.
  • the absolute value of the difference between the third characteristic and the second characteristic of the central user of each user group is calculated respectively, and the user group with the smallest absolute value of the difference is used as the target user group corresponding to the user to be returned.
  • a corresponding return visit channel is preset for each user group.
  • the return visit channel set for user group 1 and group 2 is AI voice return visit
  • the return visit channel set for user group 3 is web page return visit.
  • the cluster analysis includes:
  • C1 obtain the historical data of each user in the third database, and determine the fourth characteristic of each user based on the characteristic factor and the historical data;
  • the fourth feature is a feature obtained through dimensionality reduction processing.
  • K represents the number of user groups.
  • K is any natural number from 3 to 10, then it can be divided into 3 user groups, 4 user groups, ..., 9 user groups, Divided into 10 user groups with a total of 8 grouping results.
  • C3 calculate the contour coefficient corresponding to each kind of grouping result based on the fourth characteristic of the center user of each user group corresponding to each kind of grouping result in the described multiple grouping results;
  • S ij represents the silhouette coefficient corresponding to the j th user in the ith grouping result, represents the average distance from the fourth feature of the jth user in the i-th grouping result to the fourth features of other users in the same user group, Represents the minimum value of the average distance between the fourth feature of the jth user in the i-th grouping result and the fourth features of other user groups, S i represents the silhouette coefficient corresponding to the i-th grouping result, and n represents the total number of users.
  • the silhouette coefficient is an evaluation method for the quality of the grouping results, which reflects the cohesion and separation of the clustering method. If the inner clustering of the same cluster is higher, and the separation degree of different clusters is higher, the clustering effect is better, and the closer Si is to 1, it means The smaller the value, the better the clustering effect.
  • the preset value is 1, and the grouping result with the contour coefficient closest to 1 is used as the target grouping result.
  • S4 carry out identity verification to the user to be returned based on the target return visit channel, when identity verification passes, based on the target return visit channel and the return visit volume to the user to be returned to visit
  • the user carries out a return visit, and generates a return visit report.
  • the performing identity verification on the user to be returned based on the target return visit channel includes:
  • the return visit channel is an AI voice return visit
  • connect to the second client terminal corresponding to the identifier obtain the first audio data of the user of the second client terminal within the first preset time period, and based on the An audio data authenticates the user to be returned;
  • the return visit channel is a webpage return visit
  • connect to the third client terminal corresponding to the identifier acquire video data of the user of the third client terminal within a second preset time period, and perform a data analysis on the third client terminal based on the video data.
  • the user to be returned is authenticated.
  • the performing identity verification on the user to be returned based on the first audio data includes:
  • the performing identity verification on the user to be returned based on the video data includes:
  • Framing the video data to obtain an image sequence input the image sequence into a face recognition model to obtain the target face feature, and obtain the standard face feature corresponding to the user to be returned from the fifth database. If the similarity value between the target face feature and the standard face feature is less than the face similarity threshold, it is determined that the identity verification of the user to be returned is passed.
  • the return visit is performed to the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated, including:
  • the return visit channel is an AI voice return visit
  • record the second audio data generated by the user to be returned answering the return visit volume and convert the second audio data into text information
  • the ASR technology is used to convert the audio data Converted to text information
  • merge the text information and the return visit volume and obtain the return visit report corresponding to the user to be returned visit;
  • a return visit report is generated based on return visit information fed back by the user to be returned visit on the return visit webpage.
  • the intelligent return visit method proposed in the present application firstly determines the return visit category corresponding to the user to be returned based on user data, establishes a user portrait for the user to be returned, and generates a return visit volume for the user to be returned based on the user portrait and the return visit category.
  • this step generates a corresponding return visit volume for each user to be returned, so that the return visit volume is more relevant to the user, reduces the amount of return visit problems while meeting user needs, and can speed up return visit efficiency; then, obtain the characteristic factor corresponding to the return visit category, according to The characteristic factor obtains the first characteristic of the user to be returned, and the target return channel corresponding to the user to be returned is determined based on the first characteristic.
  • This step can intelligently determine the return channel corresponding to the user, making the return visit more flexible and more in line with user needs; finally, based on The target return visit channel performs identity verification on the returning user. When the identity verification is passed, the return visit is conducted based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • This step ensures the identity accuracy of the return visit user, and the return visit channel includes AI voice return visits and web page return visits eliminate the need for manual participation in return visits and reduce labor costs. Therefore, the present application reduces labor costs and improves return visit efficiency.
  • FIG. 2 it is a schematic diagram of a module of an intelligent return visit device according to an embodiment of the present application.
  • the smart return visit device 100 described in this application can be installed in an electronic device. According to the implemented functions, the intelligent return visit device 100 may include an analysis module 110 , a generation module 120 , a determination module 130 and a return visit module 140 .
  • the modules described in this application may also be referred to as units, which refer to a series of computer program segments that can be executed by the processor of the electronic device and can perform fixed functions, and are stored in the memory of the electronic device.
  • each module/unit is as follows:
  • the parsing module 110 is configured to parse the return visit request sent by the first client, obtain the identifier of the user to be returned carried in the return visit request, and obtain user data of the user to be returned from the first database based on the identifier.
  • the first client may be the backend server of the insurance company, or may be a client (work computer or mobile phone) corresponding to an employee of the insurance company.
  • the identifier of the user to be returned may be the ID number of the user, and the identifier of the user may be the identifier of one user or the identifiers of a group of users.
  • the user data includes the user's basic information data and the user's product information data.
  • the basic information includes name, age, address, occupation, income, etc.
  • the product information data includes the insurance type purchased by the user, insurance premium, and insurance payment. Frequency, sales channels, time of purchase, etc.
  • the generating module 120 is configured to determine the return visit category corresponding to the user to be returned based on the user data, and establish a user portrait for the user to be returned, and generate a user portrait for the user to be returned based on the user portrait and the return visit category. Back to access volume.
  • the determining the return visit category corresponding to the user to be returned based on the user data includes: determining the return visit category corresponding to the user to be returned according to the insurance category in the product information data.
  • the return visit category corresponding to user 1 is critical illness insurance return visit
  • the return visit category corresponding to user 2 is participating insurance return visit.
  • the preset keyword set is matched with the user data to obtain index data of multiple dimensions of the user, so as to establish the user portrait of the user to be returned to visit.
  • the preset keyword set may be age, income, insurance type, payment frequency
  • the user portrait corresponding to user 1 may be ⁇ 50 years old, 10,000, health insurance, payment by year ⁇
  • the user corresponding to user 2 The portrait can be ⁇ 25 years old, 8000, participating insurance, paid by year ⁇ .
  • the generating a return visit volume for the user to be returned based on the user portrait and the return visit category includes:
  • A1 extracting data corresponding to preset index items from the user data to generate a first questionnaire
  • the preset indicator items may be the user's mobile phone number and communication address. It is assumed that the mobile phone number recorded in the user data of user 1 is 135xxxxxxxx, and the communication address is No. , then the generated first questionnaire of User 1 can be "Is your mobile phone number 135xxxxxxxx? If not, please provide your new mobile phone number", "Excuse me, your mailing address is xx hometown x, xx district, Shenzhen, Guangdong province. Is the building xx number? If not, please provide a new mailing address.” Through the first questionnaire, it can be judged whether the communication information of the user to be returned has been updated.
  • A2 Obtain the second questionnaire corresponding to the return visit category from the second database, and extract a third questionnaire from the second questionnaire based on the user portrait;
  • the standard questionnaire corresponding to each type of return visit is pre-stored in the second database.
  • the content of the standard questionnaire is relatively comprehensive and includes many return visit questions.
  • the return visit questions in the second questionnaire need to be selected and processed in a targeted manner. to improve return visit efficiency.
  • user 1 may pay more attention to the compensation period and compensation ratio, so the questions related to the compensation period and compensation ratio are selected from the second questionnaire corresponding to user 1 as the third questionnaire of user 1; user 2 It is possible to pay more attention to the exemption clause and historical interest rate, then the questions related to the exemption clause and the historical interest rate are selected from the second questionnaire corresponding to user 2 as the third questionnaire of user 2.
  • the obtained return interview questionnaire is more relevant to the user, which reduces the amount of return interview questions while meeting the user's needs, and can speed up the return interview efficiency.
  • the determining module 130 is configured to obtain a characteristic factor corresponding to the return visit category, obtain a first characteristic of the user to be returned according to the characteristic factor, and determine a target return visit channel corresponding to the user to be returned based on the first characteristic.
  • This embodiment pre-sets corresponding characteristic factors for each type of return visit.
  • the characteristic factors corresponding to participating insurance may be premiums, income, historical premiums, and age.
  • the characteristic values corresponding to the above characteristic factors can be determined based on user data (for example, , if the premium is 5,000 yuan, the eigenvalue corresponding to the premium may be 5,000.
  • the value after discretization and normalization of 5,000 may be used as the eigenfactor corresponding to the premium, and the discretization and normalization
  • an array formed by eigenvalues corresponding to each eigenfactor is used as the first feature.
  • the determining of the target return visit channel corresponding to the user to be returned based on the first feature includes:
  • the PCA Principal Component Analysis
  • the PCA algorithm is used for the original feature. For all variables, delete redundant variables from the repeated variables (closely related variables), and establish as few new variables as possible, so that the new variables are not related to each other, and the new variables retain the original information as much as possible, that is, the m-dimensional feature map To n dimensions (n ⁇ m), this step can make subsequent feature processing more efficient.
  • the third feature is calculated based on the feature factor and user data of each user in the third database, and the calculation process is the same as the calculation process of the first feature, which is not repeated here.
  • the absolute value of the difference between the third characteristic and the second characteristic of the central user of each user group is calculated respectively, and the user group with the smallest absolute value of the difference is used as the target user group corresponding to the user to be returned.
  • a corresponding return visit channel is preset for each user group.
  • the return visit channel set for user group 1 and group 2 is AI voice return visit
  • the return visit channel set for user group 3 is web page return visit.
  • the cluster analysis includes:
  • C1 obtain the historical data of each user in the third database, and determine the fourth characteristic of each user based on the characteristic factor and the historical data;
  • the fourth feature is a feature obtained through dimensionality reduction processing.
  • K represents the number of user groups.
  • K is any natural number from 3 to 10, then it can be divided into 3 user groups, 4 user groups, ..., 9 user groups, Divided into 10 user groups with a total of 8 grouping results.
  • C3 calculate the contour coefficient corresponding to each kind of grouping result based on the fourth characteristic of the center user of each user group corresponding to each kind of grouping result in the described multiple grouping results;
  • S ij represents the silhouette coefficient corresponding to the j th user in the ith grouping result, represents the average distance from the fourth feature of the jth user in the i-th grouping result to the fourth features of other users in the same user group, Represents the minimum value of the average distance between the fourth feature of the jth user in the i-th grouping result and the fourth features of other user groups, S i represents the silhouette coefficient corresponding to the i-th grouping result, and n represents the total number of users.
  • the silhouette coefficient is an evaluation method for the quality of the grouping results, which reflects the cohesion and separation of the clustering method. If the inner clustering of the same cluster is higher, and the separation degree of different clusters is higher, the clustering effect is better, and the closer Si is to 1, it means The smaller the value, the better the clustering effect.
  • the preset value is 1, and the grouping result with the contour coefficient closest to 1 is used as the target grouping result.
  • the return visit module 140 is configured to perform identity verification on the user to be returned based on the target return visit channel, and when the identity verification is passed, conduct return visit to the user to be returned based on the target return visit channel and the return visit volume, and generate Return visit report.
  • the performing identity verification on the user to be returned based on the target return visit channel includes:
  • the return visit channel is an AI voice return visit
  • connect to the second client terminal corresponding to the identifier obtain the first audio data of the user of the second client terminal within the first preset time period, and based on the first The audio data authenticates the user to be returned;
  • the return visit channel is a webpage return visit
  • connect to the third client terminal corresponding to the identifier acquire video data of the user of the third client terminal within a second preset time period, and perform a data analysis on the third client terminal based on the video data.
  • the user to be returned is authenticated.
  • the performing identity verification on the user to be returned based on the first audio data includes:
  • the performing identity verification on the user to be returned based on the video data includes:
  • Framing the video data to obtain an image sequence input the image sequence into a face recognition model to obtain the target face feature, and obtain the standard face feature corresponding to the user to be returned from the fifth database. If the similarity value between the target face feature and the standard face feature is less than the face similarity threshold, it is determined that the identity verification of the user to be returned is passed.
  • the return visit is performed to the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated, including:
  • the return visit channel is an AI voice return visit
  • record the second audio data generated by the user to be returned answering the return visit volume and convert the second audio data into text information
  • the ASR technology is used to convert the audio data is converted into text information
  • merge the text information and the return visit volume and obtain the return visit report corresponding to the user to be returned visit;
  • a return visit report is generated based on return visit information fed back by the user to be returned visit on the return visit webpage.
  • FIG. 3 it is a schematic structural diagram of an electronic device for implementing an intelligent return visit method according to an embodiment of the present application.
  • the electronic device 1 is a device that can automatically perform numerical calculation and/or information processing according to pre-set or stored instructions.
  • the electronic device 1 may be a computer, a single network server, a server group composed of multiple network servers, or a cloud based on cloud computing composed of a large number of hosts or network servers, wherein cloud computing is a kind of distributed computing, A super virtual computer consisting of a collection of loosely coupled computers.
  • the electronic device 1 includes, but is not limited to, a memory 11, a processor 12, and a network interface 13 that can be communicatively connected to each other through a system bus.
  • the memory 11 stores an intelligent return visit program 10, and the intelligent return visit program 10 is executable by the processor 12 .
  • FIG. 3 only shows the electronic device 1 having the components 11-13 and the smart return visit program 10. Those skilled in the art can understand that the structure shown in FIG. 3 does not constitute a limitation on the electronic device 1, and may include a Fewer or more components are shown, or some components are combined, or a different arrangement of components.
  • the memory 11 includes a memory and at least one type of readable storage medium.
  • the memory provides a cache for the operation of the electronic device 1;
  • the readable storage medium can be, for example, flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM) ), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable read only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. non-volatile storage media.
  • the readable storage medium may be an internal storage unit of the electronic device 1, such as a hard disk of the electronic device 1; in other embodiments, the non-volatile storage medium may also be an external storage unit of the electronic device 1
  • a storage device such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash memory card (Flash Card), etc. equipped on the electronic device 1.
  • the readable storage medium of the memory 11 is generally used to store the operating system and various application software installed in the electronic device 1 , for example, to store the code of the smart return program 10 in an embodiment of the present application.
  • the memory 11 can also be used to temporarily store various types of data that have been output or will be output.
  • the processor 12 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips.
  • the processor 12 is generally used to control the overall operation of the electronic device 1, such as performing control and processing related to data interaction or communication with other devices.
  • the processor 12 is configured to run the program code or process data stored in the memory 11 , for example, run the intelligent return access program 10 and the like.
  • the network interface 13 may include a wireless network interface or a wired network interface, and the network interface 13 is used to establish a communication connection between the electronic device 1 and a client (not shown in the figure).
  • the electronic device 1 may further include a user interface, and the user interface may include a display (Display), an input unit such as a keyboard (Keyboard), and an optional user interface may also include a standard wired interface and a wireless interface.
  • the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, and the like.
  • the display may also be appropriately called a display screen or a display unit, which is used for displaying information processed in the electronic device 1 and for displaying a visualized user interface.
  • the smart return program 10 stored in the memory 11 in the electronic device 1 is a combination of multiple instructions, and when running in the processor 12, it can realize:
  • the characteristic factor corresponding to the return visit category obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • the identity verification is performed on the user to be returned based on the target return visit channel, and when the identity verification is passed, the return visit is performed on the user to be returned based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • the above-mentioned intelligent return visit program 10 by the processor 12, reference may be made to the description of the relevant steps in the corresponding embodiment of FIG. 1, and details are not described herein. It should be emphasized that, in order to further ensure the privacy and security of the above-mentioned user data, the above-mentioned user data can also be stored in a node of a blockchain.
  • the modules/units integrated in the electronic device 1 may be stored in a computer-readable storage medium.
  • the computer-readable storage medium may be non-volatile or non-volatile.
  • the computer-readable storage medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory) ).
  • the computer-readable storage medium stores an intelligent return visit program 10, and the intelligent return visit program 10 can be executed by one or more processors to realize the following steps:
  • the characteristic factor corresponding to the return visit category obtain the first characteristic of the user to be returned according to the characteristic factor, and determine the target return visit channel corresponding to the user to be returned based on the first characteristic;
  • the user to be returned is authenticated based on the target return visit channel, and when the identity verification is passed, the user to be returned visit is returned to the user based on the target return visit channel and the return visit volume, and a return visit report is generated.
  • modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment.
  • each functional module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.
  • the above-mentioned integrated units can be implemented in the form of hardware, or can be implemented in the form of hardware plus software function modules.
  • the blockchain referred to in this application is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Finance (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Data Mining & Analysis (AREA)
  • Development Economics (AREA)
  • Accounting & Taxation (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Computer Security & Cryptography (AREA)
  • Marketing (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种人工智能技术和数据分析,揭露一种智能回访方法,包括:基于用户数据确定待回访用户对应的回访类别,并为待回访用户建立用户画像,基于用户画像及回访类别为待回访用户生成回访问卷;获取回访类别对应的特征因子,根据特征因子得到待回访用户的第一特征,基于第一特征确定待回访用户对应的目标回访渠道;基于目标回访渠道对待回访用户进行身份验证,当身份验证通过时,基于目标回访渠道及回访问卷对待回访用户进行回访,并生成回访报告。该方法实现了减少人力成本、提升回访效率。

Description

智能回访方法、装置、电子设备及可读存储介质
本申请要求于2020年11月19日提交中国专利局、申请号为CN202011305616.1、名称为“智能回访方法、装置、电子设备及可读存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能技术领域和数据分析领域,尤其涉及一种智能回访方法、装置、电子设备及可读存储介质。
背景技术
随着科技的发展,产品和服务越来越多样化,为提升用户体验,改进产品和服务的质量,通常通过对用户进行回访以了解用户需求。发明人意识到,当前的回访方式通常是人工整理好回访问题,再对用户进行电话回访,这种回访方式人力成本大、回访效率低、回访渠道单一。因此,亟需一种智能回访方法,以减少人力成本、提升回访效率。
发明内容
本申请提供的智能回访方法,包括:
解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
本申请还提供一种智能回访装置,所述装置包括:
解析模块,用于解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
生成模块,用于基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
确定模块,用于获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
回访模块,用于基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
本申请还提供一种电子设备,所述电子设备包括:
至少一个处理器;以及,
与所述至少一个处理器通信连接的存储器;其中,
所述存储器存储有可被所述至少一个处理器执行的智能回访程序,所述智能回访程序被所述至少一个处理器执行,以使所述至少一个处理器能够执行如下步骤:
解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
本申请还提供一种计算机可读存储介质,所述计算机可读存储介质上存储有智能回访程序,所述智能回访程序可被一个或者多个处理器执行,以实现如下步骤:
解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
附图说明
图1为本申请一实施例提供的智能回访方法的流程示意图;
图2为本申请一实施例提供的智能回访装置的模块示意图;
图3为本申请一实施例提供的实现智能回访方法的电子设备的结构示意图;
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本申请进行进一步详细说明。应当理解,此处所描述的具体实施例仅用以解释本申请,并不用于限定本申请。基于本申请中的实施例,本领域普通技术人员在没有做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。
需要说明的是,在本申请中涉及“第一”、“第二”等的描述仅用于描述目的,而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此,限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。另外,各个实施例之间的技术方案可以相互结合,但是必须是以本领域普通技术人员能够实现为基础,当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在,也不在本申请要求的保护范围之内。
本申请实施例可以基于人工智能技术对相关的数据进行获取和处理。其中,人工智能(Artificial Intelligence,AI)是利用数字计算机或者数字计算机控制的机器模拟、延伸和扩展人的智能,感知环境、获取知识并使用知识获得最佳结果的理论、方法、技术及应用系统。
人工智能基础技术一般包括如传感器、专用人工智能芯片、云计算、分布式存储、大数据处理技术、操作/交互系统、机电一体化等技术。人工智能软件技术主要包括计算机视觉技术、机器人技术、生物识别技术、语音处理技术、自然语言处理技术以及机器学习/深度学习等几大方向。
本申请提供一种智能回访方法。参照图1所示,为本申请一实施例提供的智能回访方法的流程示意图。该方法可以由一个电子设备执行,该电子设备可以由软件和/或硬件实现。
本实施例中,智能回访方法包括:
S1、解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据。
本实施例以保险公司为保险用户进行回访为例进行说明,所述第一客户端可以是保险 公司的后台服务器,也可以是保险公司员工对应的客户端(工作电脑或手机)。
所述待回访用户的标识可以是用户的身份证号码,所述用户的标识可以是一个用户的标识,也可以是一批用户的标识。所述用户数据包括用户的基本信息数据及用户的产品信息数据,所述基本信息包括姓名、年龄、住址、职业、收入等,所述产品信息数据包括用户购买的险种类别、险种保费、险种缴费频率、销售渠道、购买时间等。
S2、基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷。
本实施例中,所述基于所述用户数据确定所述待回访用户对应的回访类别包括:根据产品信息数据中险种类别确定待回访用户对应的回访类别。
例如,若用户1购买了重疾险,则用户1对应的回访类别为重疾险回访,若用户2购买了分红险,则用户2对应的回访类别为分红险回访。
本实施例将预设关键词集合与用户数据进行匹配得到用户的多个维度的指标数据,以建立待回访用户的用户画像,具体建立用户画像的过程可参考现有技术,在此不再赘述。例如,所述预设关键词集合可以是年龄、收入、险种类别、缴费频率,用户1对应的用户画像可以为{50岁,1万,健康险,以年次缴费},用户2对应的用户画像可以为{25岁,8000,分红险,以年次缴费}。
所述基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷包括:
A1、从所述用户数据中抽取预设指标项对应的数据生成第一问卷;
本实施例中,所述预设指标项可以是用户的手机号码、通讯地址,假设用户1的用户数据中记录的手机号码是135xxxxxxxx、通讯地址是广东省深圳市xx区xx家园x栋xx号,则生成的用户1的第一问卷可以是“请问您的手机号码是135xxxxxxxx吗?若不是,请提供您的新手机号码”、“请问您的通讯地址是广东省深圳市xx区xx家园x栋xx号吗?若不是,请提供新的通讯地址”,通过第一问卷可判断待回访用户的通讯信息是否有更新。
A2、从第二数据库中获取所述回访类别对应的第二问卷,基于所述用户画像从所述第二问卷中抽取第三问卷;
所述第二数据库中预先存储有每种回访类别对应的标准问卷,所述标准问卷的内容较为全面,所包括的回访问题较多,需对第二问卷中的回访问题进行针对性抽取处理,以提高回访效率。
例如,根据用户画像可推导,用户1可能更为关注赔付周期、赔付比例,则从用户1对应的第二问卷中抽取与赔付周期、赔付比例相关的问题作为用户1的第三问卷;用户2可能更为关注免责条款、历史利率,则从用户2对应的第二问卷中抽取与免责条款、历史利率相关的问题作为用户2的第三问卷。
A3、合并所述第一问卷及第三问卷,得到所述待回访用户对应的回访问卷。
该步骤通过从第二问卷中抽取第三问卷,并将第一问卷与第三问卷合并,使得得到的回访问卷与用户更相关,在满足用户需求的同时减少回访问题量,可加快回访效率。
S3、获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道。
本实施例预先为每种回访类别设置了对应的特征因子,例如,分红险对应的特征因子可以是保费、收入、历史保费、年龄,基于用户数据可确定以上各个特征因子对应的特征值(例如,若保费为5000元,则保费对应的特征值可以是5000,在其他实施例中,也可以将5000进行离散化、归一化处理后的值作为保费对应的特征因子,离散化、归一化处理可参考现有技术,在此不做赘述),将各个特征因子对应的特征值所组成的数组作为第一特征。
所述基于所述第一特征确定所述待回访用户对应的目标回访渠道包括:
B1、对所述第一特征执行降维处理,得到第二特征;
本实施例采用PCA(Principal Component Analysis,主成分分析)算法对第一特征执行降维处理,当两个变量有相关性时,认为两个变量有一定的信息重叠,PCA算法是对于原有的所有变量,从重复的变量(关系紧密的变量)中删去多余变量,建立尽可能少的新变量,使得新变量两两不相关,且新变量尽可能保留原有信息,即将m维特征映射到n维上(n<m),本步骤可使后续的特征处理的效率更高。
B2、获取第三数据库中已执行聚类分析的各个用户组的中心用户的第三特征;
所述第三特征是基于所述特征因子及第三数据库中各个用户的用户数据计算得到的,其计算过程与第一特征的计算过程相同,在此不再赘述。
B3、基于所述第二特征及第三特征确定所述待回访用户对应的目标用户组;
本实施例中,分别计算各个用户组的中心用户的第三特征与所述第二特征的差值绝对值,将差值绝对值最小的用户组作为所述待回访用户对应的目标用户组。
B4、基于所述目标用户组及用户组与回访渠道对应的映射关系确定所述待回访用户对应的目标回访渠道。
本实施例预先为每个用户组设置了对应的回访渠道,例如,为用户组1、用于组2设置的回访渠道为AI语音回访,为用户组3设置的回访渠道为网页回访。
所述聚类分析包括:
C1、获取第三数据库中各个用户的历史数据,基于所述特征因子及所述历史数据确定各个用户的第四特征;
所述第四特征为经过降维处理得到的特征。
C2、基于所述第四特征及K均值聚类算法对所述第三数据库中的用户进行分组,其中,K分别取值为预设数值范围内的各个自然数,K的一个取值对应一种分组结果,得到多种分组结果;
K表示用户组的数量,本实施例中,K为3~10中的任一个自然数,则可以得到分为3个用户组、分为4个用户组、……、分为9个用户组、分为10个用户组共8种分组结果。
以K=3举例说明用户分组过程:任取3个用户的第四特征作为三个初始聚类中心,然后计算剩余用户与各个聚类中心之间的第四特征的距离,把每个用户分配给距离它最近的聚类,每分配一个用户,聚类的聚类中心会根据聚类中现有的用户被重新计算,如此循环直至将所有用户分组完成。
C3、基于所述多种分组结果中每种分组结果对应的各个用户组的中心用户的第四特征计算每种分组结果对应的轮廓系数;
所述轮廓系数对应的计算公式为:
Figure PCTCN2021123905-appb-000001
Figure PCTCN2021123905-appb-000002
其中,S ij表示第i种分组结果中第j个用户对应的轮廓系数,
Figure PCTCN2021123905-appb-000003
表示第i种分组结果中第j个用户的第四特征到同一个用户组中其他用户的第四特征的平均距离,
Figure PCTCN2021123905-appb-000004
表示第i种分组结果中第j个用户的第四特征到其他用户组的第四特征的平均距离的最小值,S i表示第i种分组结果对应的轮廓系数,n表示用户的总数量。
轮廓系数是分组结果好坏的一种评价方式,反映了该聚类方法的内聚度和分离度。若同一个簇的内聚类越高,不同簇的分离度越高,则聚类效果越好,S i越接近1表示
Figure PCTCN2021123905-appb-000005
越小,聚类效果越好。
C4、将轮廓系数最接近预设数值的分组结果作为目标分组结果。
本实施例中,预设数值为1,将轮廓系数最接近1的分组结果作为目标分组结果。
S4、基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基 于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
所述基于所述目标回访渠道对所述待回访用户进行身份验证包括:
D1、当所述回访渠道为AI语音回访时,连接所述标识对应的第二客户端,获取所述第二客户端的用户在第一预设时间段内的第一音频数据,基于所述第一音频数据对所述待回访用户进行身份验证;
D2、若所述回访渠道为网页回访,连接所述标识对应的第三客户端,获取所述第三客户端的用户在第二预设时间段内的视频数据,基于所述视频数据对所述待回访用户进行身份验证。
所述基于所述第一音频数据对所述待回访用户进行身份验证包括:
对所述第一音频数据进行短时傅里叶变换和/或短时傅里叶逆变换,得到所述第二客户端的用户的时域信号数据,将所述时域信号数据输入声纹识别模型,得到目标声纹特征,从第四数据库中获取所述待回访用户对应的标准声纹特征,若所述目标声纹特征与标准声纹特征的相似度值小于声纹相似度阈值,则判断所述待回访用户身份验证通过;
所述基于所述视频数据对所述待回访用户进行身份验证包括:
对所述视频数据进行分帧,得到图像序列,将所述图像序列输入人脸识别模型,得到目标人脸特征,从第五数据库中获取所述待回访用户对应的标准人脸特征,若所述目标人脸特征与标准人脸特征的相似度值小于人脸相似度阈值,则判断所述待回访用户身份验证通过。
所述基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告,包括:
若所述回访渠道为AI语音回访,录制所述待回访用户回答所述回访问卷所产生的第二音频数据,将所述第二音频数据转换为文本信息(本实施例采用ASR技术将音频数据转换为文本信息),合并所述文本信息及所述回访问卷,得到所述待回访用户对应的回访报告;
若所述回访渠道为网页回访,基于所述待回访用户在回访网页上反馈的回访信息生成回访报告。
由上述实施例可知,本申请提出的智能回访方法,首先,基于用户数据确定待回访用户对应的回访类别,并为待回访用户建立用户画像,基于用户画像及回访类别为待回访用户生成回访问卷,该步骤为每个待回访用户生成对应的回访问卷,使得回访问卷与用户更相关,在满足用户需求的同时减少回访问题量,可加快回访效率;接着,获取回访类别对应的特征因子,根据特征因子得到待回访用户的第一特征,基于第一特征确定待回访用户对应的目标回访渠道,本步骤可智能确定用户对应的回访渠道,使得回访更为灵活,更符合用户需求;最后,基于目标回访渠道对待回访用户进行身份验证,当身份验证通过时,基于目标回访渠道及回访问卷对待回访用户进行回访,并生成回访报告,该步骤保证了待回访用户的身份准确性,且回访渠道包括AI语音回访及网页回访,从而不需人工参与回访,减少了人力成本。因此,本申请减少了人力成本、提升了回访效率。
如图2所示,为本申请一实施例提供的智能回访装置的模块示意图。
本申请所述智能回访装置100可以安装于电子设备中。根据实现的功能,所述智能回访装置100可以包括解析模块110、生成模块120、确定模块130及回访模块140。本申请所述模块也可以称之为单元,是指一种能够被电子设备处理器所执行,并且能够完成固定功能的一系列计算机程序段,其存储在电子设备的存储器中。
在本实施例中,关于各模块/单元的功能如下:
解析模块110,用于解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据。
本实施例以保险公司为保险用户进行回访为例进行说明,所述第一客户端可以是保险 公司的后台服务器,也可以是保险公司员工对应的客户端(工作电脑或手机)。
所述待回访用户的标识可以是用户的身份证号码,所述用户的标识可以是一个用户的标识,也可以是一批用户的标识。所述用户数据包括用户的基本信息数据及用户的产品信息数据,所述基本信息包括姓名、年龄、住址、职业、收入等,所述产品信息数据包括用户购买的险种类别、险种保费、险种缴费频率、销售渠道、购买时间等。
生成模块120,用于基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷。
本实施例中,所述基于所述用户数据确定所述待回访用户对应的回访类别包括:根据产品信息数据中险种类别确定待回访用户对应的回访类别。
例如,若用户1购买了重疾险,则用户1对应的回访类别为重疾险回访,若用户2购买了分红险,则用户2对应的回访类别为分红险回访。
本实施例将预设关键词集合与用户数据进行匹配得到用户的多个维度的指标数据,以建立待回访用户的用户画像,具体建立用户画像的过程可参考现有技术,在此不再赘述。例如,所述预设关键词集合可以是年龄、收入、险种类别、缴费频率,用户1对应的用户画像可以为{50岁,1万,健康险,以年次缴费},用户2对应的用户画像可以为{25岁,8000,分红险,以年次缴费}。
所述基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷包括:
A1、从所述用户数据中抽取预设指标项对应的数据生成第一问卷;
本实施例中,所述预设指标项可以是用户的手机号码、通讯地址,假设用户1的用户数据中记录的手机号码是135xxxxxxxx、通讯地址是广东省深圳市xx区xx家园x栋xx号,则生成的用户1的第一问卷可以是“请问您的手机号码是135xxxxxxxx吗?若不是,请提供您的新手机号码”、“请问您的通讯地址是广东省深圳市xx区xx家园x栋xx号吗?若不是,请提供新的通讯地址”,通过第一问卷可判断待回访用户的通讯信息是否有更新。
A2、从第二数据库中获取所述回访类别对应的第二问卷,基于所述用户画像从所述第二问卷中抽取第三问卷;
所述第二数据库中预先存储有每种回访类别对应的标准问卷,所述标准问卷的内容较为全面,所包括的回访问题较多,需对第二问卷中的回访问题进行针对性抽取处理,以提高回访效率。
例如,根据用户画像可推导,用户1可能更为关注赔付周期、赔付比例,则从用户1对应的第二问卷中抽取与赔付周期、赔付比例相关的问题作为用户1的第三问卷;用户2可能更为关注免责条款、历史利率,则从用户2对应的第二问卷中抽取与免责条款、历史利率相关的问题作为用户2的第三问卷。
A3、合并所述第一问卷及第三问卷,得到所述待回访用户对应的回访问卷。
该步骤通过从第二问卷中抽取第三问卷,并将第一问卷与第三问卷合并,使得得到的回访问卷与用户更相关,在满足用户需求的同时减少回访问题量,可加快回访效率。
确定模块130,用于获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道。
本实施例预先为每种回访类别设置了对应的特征因子,例如,分红险对应的特征因子可以是保费、收入、历史保费、年龄,基于用户数据可确定以上各个特征因子对应的特征值(例如,若保费为5000元,则保费对应的特征值可以是5000,在其他实施例中,也可以将5000进行离散化、归一化处理后的值作为保费对应的特征因子,离散化、归一化处理可参考现有技术,在此不做赘述),将各个特征因子对应的特征值所组成的数组作为第一特征。
所述基于所述第一特征确定所述待回访用户对应的目标回访渠道包括:
B1、对所述第一特征执行降维处理,得到第二特征;
本实施例采用PCA(Principal Component Analysis,主成分分析)算法对第一特征执行降维处理,当两个变量有相关性时,认为两个变量有一定的信息重叠,PCA算法是对于原有的所有变量,从重复的变量(关系紧密的变量)中删去多余变量,建立尽可能少的新变量,使得新变量两两不相关,且新变量尽可能保留原有信息,即将m维特征映射到n维上(n<m),本步骤可使后续的特征处理的效率更高。
B2、获取第三数据库中已执行聚类分析的各个用户组的中心用户的第三特征;
所述第三特征是基于所述特征因子及第三数据库中各个用户的用户数据计算得到的,其计算过程与第一特征的计算过程相同,在此不再赘述。
B3、基于所述第二特征及第三特征确定所述待回访用户对应的目标用户组;
本实施例中,分别计算各个用户组的中心用户的第三特征与所述第二特征的差值绝对值,将差值绝对值最小的用户组作为所述待回访用户对应的目标用户组。
B4、基于所述目标用户组及用户组与回访渠道对应的映射关系确定所述待回访用户对应的目标回访渠道。
本实施例预先为每个用户组设置了对应的回访渠道,例如,为用户组1、用于组2设置的回访渠道为AI语音回访,为用户组3设置的回访渠道为网页回访。
所述聚类分析包括:
C1、获取第三数据库中各个用户的历史数据,基于所述特征因子及所述历史数据确定各个用户的第四特征;
所述第四特征为经过降维处理得到的特征。
C2、基于所述第四特征及K均值聚类算法对所述第三数据库中的用户进行分组,其中,K分别取值为预设数值范围内的各个自然数,K的一个取值对应一种分组结果,得到多种分组结果;
K表示用户组的数量,本实施例中,K为3~10中的任一个自然数,则可以得到分为3个用户组、分为4个用户组、……、分为9个用户组、分为10个用户组共8种分组结果。
以K=3举例说明用户分组过程:任取3个用户的第四特征作为三个初始聚类中心,然后计算剩余用户与各个聚类中心之间的第四特征的距离,把每个用户分配给距离它最近的聚类,每分配一个用户,聚类的聚类中心会根据聚类中现有的用户被重新计算,如此循环直至将所有用户分组完成。
C3、基于所述多种分组结果中每种分组结果对应的各个用户组的中心用户的第四特征计算每种分组结果对应的轮廓系数;
所述轮廓系数对应的计算公式为:
Figure PCTCN2021123905-appb-000006
Figure PCTCN2021123905-appb-000007
其中,S ij表示第i种分组结果中第j个用户对应的轮廓系数,
Figure PCTCN2021123905-appb-000008
表示第i种分组结果中第j个用户的第四特征到同一个用户组中其他用户的第四特征的平均距离,
Figure PCTCN2021123905-appb-000009
表示第i种分组结果中第j个用户的第四特征到其他用户组的第四特征的平均距离的最小值,S i表示第i种分组结果对应的轮廓系数,n表示用户的总数量。
轮廓系数是分组结果好坏的一种评价方式,反映了该聚类方法的内聚度和分离度。若同一个簇的内聚类越高,不同簇的分离度越高,则聚类效果越好,S i越接近1表示
Figure PCTCN2021123905-appb-000010
越小,聚类效果越好。
C4、将轮廓系数最接近预设数值的分组结果作为目标分组结果。
本实施例中,预设数值为1,将轮廓系数最接近1的分组结果作为目标分组结果。
回访模块140,用于基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
所述基于所述目标回访渠道对所述待回访用户进行身份验证包括:
D1、若所述回访渠道为AI语音回访,连接所述标识对应的第二客户端,获取所述第二客户端的用户在第一预设时间段内的第一音频数据,基于所述第一音频数据对所述待回访用户进行身份验证;
D2、若所述回访渠道为网页回访,连接所述标识对应的第三客户端,获取所述第三客户端的用户在第二预设时间段内的视频数据,基于所述视频数据对所述待回访用户进行身份验证。
所述基于所述第一音频数据对所述待回访用户进行身份验证包括:
对所述第一音频数据进行短时傅里叶变换和/或短时傅里叶逆变换,得到所述第二客户端的用户的时域信号数据,将所述时域信号数据输入声纹识别模型,得到目标声纹特征,从第四数据库中获取所述待回访用户对应的标准声纹特征,若所述目标声纹特征与标准声纹特征的相似度值小于声纹相似度阈值,则判断所述待回访用户身份验证通过;
所述基于所述视频数据对所述待回访用户进行身份验证包括:
对所述视频数据进行分帧,得到图像序列,将所述图像序列输入人脸识别模型,得到目标人脸特征,从第五数据库中获取所述待回访用户对应的标准人脸特征,若所述目标人脸特征与标准人脸特征的相似度值小于人脸相似度阈值,则判断所述待回访用户身份验证通过。
所述基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告,包括:
当所述回访渠道为AI语音回访时,录制所述待回访用户回答所述回访问卷所产生的第二音频数据,将所述第二音频数据转换为文本信息(本实施例采用ASR技术将音频数据转换为文本信息),合并所述文本信息及所述回访问卷,得到所述待回访用户对应的回访报告;
若所述回访渠道为网页回访,基于所述待回访用户在回访网页上反馈的回访信息生成回访报告。
如图3所示,为本申请一实施例提供的实现智能回访方法的电子设备的结构示意图。
所述电子设备1是一种能够按照事先设定或者存储的指令,自动进行数值计算和/或信息处理的设备。所述电子设备1可以是计算机、也可以是单个网络服务器、多个网络服务器组成的服务器组或者基于云计算的由大量主机或者网络服务器构成的云,其中云计算是分布式计算的一种,由一群松散耦合的计算机集组成的一个超级虚拟计算机。
在本实施例中,电子设备1包括,但不仅限于,可通过系统总线相互通信连接的存储器11、处理器12、网络接口13,该存储器11中存储有智能回访程序10,所述智能回访程序10可被所述处理器12执行。图3仅示出了具有组件11-13以及智能回访程序10的电子设备1,本领域技术人员可以理解的是,图3示出的结构并不构成对电子设备1的限定,可以包括比图示更少或者更多的部件,或者组合某些部件,或者不同的部件布置。
其中,存储器11包括内存及至少一种类型的可读存储介质。内存为电子设备1的运行提供缓存;可读存储介质可为如闪存、硬盘、多媒体卡、卡型存储器(例如,SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘等的非易失性存储介质。在一些实施例中,可读存储介质可以是电子设备1的内部存储单元,例如该电子设备1的硬盘;在另一些实施例中,该非易失性存储介质也可以是电子设备1的外部存储设备,例如电子设备1上配备的插接式硬盘,智能存储卡(Smart  Media Card,SMC),安全数字(Secure Digital,SD)卡,闪存卡(Flash Card)等。本实施例中,存储器11的可读存储介质通常用于存储安装于电子设备1的操作系统和各类应用软件,例如存储本申请一实施例中的智能回访程序10的代码等。此外,存储器11还可以用于暂时地存储已经输出或者将要输出的各类数据。
处理器12在一些实施例中可以是中央处理器(Central Processing Unit,CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器12通常用于控制所述电子设备1的总体操作,例如执行与其他设备进行数据交互或者通信相关的控制和处理等。本实施例中,所述处理器12用于运行所述存储器11中存储的程序代码或者处理数据,例如运行智能回访程序10等。
网络接口13可包括无线网络接口或有线网络接口,该网络接口13用于在所述电子设备1与客户端(图中未画出)之间建立通信连接。
可选的,所述电子设备1还可以包括用户接口,用户接口可以包括显示器(Display)、输入单元比如键盘(Keyboard),可选的用户接口还可以包括标准的有线接口、无线接口。可选的,在一些实施例中,显示器可以是LED显示器、液晶显示器、触控式液晶显示器以及OLED(Organic Light-Emitting Diode,有机发光二极管)触摸器等。其中,显示器也可以适当的称为显示屏或显示单元,用于显示在电子设备1中处理的信息以及用于显示可视化的用户界面。
应该了解,所述实施例仅为说明之用,在专利申请范围上并不受此结构的限制。
所述电子设备1中的所述存储器11存储的智能回访程序10是多个指令的组合,在所述处理器12中运行时,可以实现:
解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
具体地,所述处理器12对上述智能回访程序10的具体实现方法可参考图1对应实施例中相关步骤的描述,在此不赘述。需要强调的是,为进一步保证上述用户数据的私密和安全性,上述用户数据还可以存储于一区块链的节点中。
进一步地,所述电子设备1集成的模块/单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读存储介质中。所述计算机可读存储介质可以是非易失性的,也可以是非易失性的。所述计算机可读存储介质可以包括:能够携带所述计算机程序代码的任何实体或装置、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)。
所述计算机可读存储介质上存储有智能回访程序10,所述智能回访程序10可被一个或者多个处理器执行,以实现如下步骤:
解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所 述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
在本申请所提供的几个实施例中,应该理解到,所揭露的设备,装置和方法,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的,例如,所述模块的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。
所述作为分离部件说明的模块可以是或者也可以不是物理上分开的,作为模块显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
另外,在本申请各个实施例中的各功能模块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,也可以采用硬件加软件功能模块的形式实现。
对于本领域技术人员而言,显然本申请不限于上述示范性实施例的细节,而且在不背离本申请的精神或基本特征的情况下,能够以其他的具体形式实现本申请。
因此,无论从哪一点来看,均应将实施例看作是示范性的,而且是非限制性的,本申请的范围由所附权利要求而不是上述说明限定,因此旨在将落在权利要求的等同要件的含义和范围内的所有变化涵括在本申请内。不应将权利要求中的任何附关联图标记视为限制所涉及的权利要求。
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
此外,显然“包括”一词不排除其他单元或步骤,单数不排除复数。系统权利要求中陈述的多个单元或装置也可以由一个单元或装置通过软件或者硬件来实现。第二等词语用来表示名称,而并不表示任何特定的顺序。
最后应说明的是,以上实施例仅用以说明本申请的技术方案而非限制,尽管参照较佳实施例对本申请进行了详细说明,本领域的普通技术人员应当理解,可以对本申请的技术方案进行修改或等同替换,而不脱离本申请技术方案的精神和范围。

Claims (20)

  1. 一种智能回访方法,其中,所述方法包括:
    解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
    基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
    获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
    基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
  2. 如权利要求1所述的智能回访方法,其中,所述基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷包括:
    从所述用户数据中抽取预设指标项对应的数据生成第一问卷;
    从第二数据库中获取所述回访类别对应的第二问卷,基于所述用户画像从所述第二问卷中抽取第三问卷;
    合并所述第一问卷及第三问卷,得到所述待回访用户对应的回访问卷。
  3. 如权利要求1所述的智能回访方法,其中,所述基于所述第一特征确定所述待回访用户对应的目标回访渠道包括:
    对所述第一特征执行降维处理,得到第二特征;
    获取第三数据库中已执行聚类分析的各个用户组的中心用户的第三特征;
    基于所述第二特征及第三特征确定所述待回访用户对应的目标用户组;
    基于所述目标用户组及用户组与回访渠道对应的映射关系确定所述待回访用户对应的目标回访渠道。
  4. 如权利要求3所述的智能回访方法,其中,所述聚类分析包括:
    获取第三数据库中各个用户的历史数据,基于所述特征因子及所述历史数据确定各个用户的第四特征;
    基于所述第四特征及K均值聚类算法对所述第三数据库中的用户进行分组,其中,K分别取值为预设数值范围内的各个自然数,K的一个取值对应一种分组结果,得到多种分组结果;
    基于所述多种分组结果中每种分组结果对应的各个用户组的中心用户的第四特征计算每种分组结果对应的轮廓系数;
    将轮廓系数最接近预设数值的分组结果作为目标分组结果。
  5. 如权利要求1所述的智能回访方法,其中,所述目标回访渠道包括AI语音回访及网页回访,所述基于所述目标回访渠道对所述待回访用户进行身份验证包括:
    若所述回访渠道为AI语音回访,连接所述标识对应的第二客户端,获取所述第二客户端的用户在第一预设时间段内的第一音频数据,基于所述第一音频数据对所述待回访用户进行身份验证;
    若所述回访渠道为网页回访,连接所述标识对应的第三客户端,获取所述第三客户端的用户在第二预设时间段内的视频数据,基于所述视频数据对所述待回访用户进行身份验证。
  6. 如权利要求5所述的智能回访方法,其中,所述基于所述第一音频数据对所述待回访用户进行身份验证包括:
    对所述第一音频数据进行短时傅里叶变换和/或短时傅里叶逆变换,得到所述第二客户端的用户的时域信号数据,将所述时域信号数据输入声纹识别模型,得到目标声纹特征, 从第四数据库中获取所述待回访用户对应的标准声纹特征,若所述目标声纹特征与标准声纹特征的相似度值小于声纹相似度阈值,则判断所述待回访用户身份验证通过;
    所述基于所述视频数据对所述待回访用户进行身份验证包括:
    对所述视频数据进行分帧,得到图像序列,将所述图像序列输入人脸识别模型,得到目标人脸特征,从第五数据库中获取所述待回访用户对应的标准人脸特征,若所述目标人脸特征与标准人脸特征的相似度值小于人脸相似度阈值,则判断所述待回访用户身份验证通过。
  7. 如权利要求4所述的智能回访方法,其中,所述轮廓系数对应的计算公式为:
    Figure PCTCN2021123905-appb-100001
    Figure PCTCN2021123905-appb-100002
    其中,S ij表示第i种分组结果中第j个用户对应的轮廓系数,
    Figure PCTCN2021123905-appb-100003
    表示第i种分组结果中第j个用户的第四特征到同一个用户组中其他用户的第四特征的平均距离,
    Figure PCTCN2021123905-appb-100004
    表示第i种分组结果中第j个用户的第四特征到其他用户组的第四特征的平均距离的最小值,S i表示第i种分组结果对应的轮廓系数,n表示用户的总数量。
  8. 一种智能回访装置,其中,所述装置包括:
    解析模块,用于解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
    生成模块,用于基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
    确定模块,用于获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
    回访模块,用于基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
  9. 一种电子设备,其中,所述电子设备包括:
    至少一个处理器;以及,
    与所述至少一个处理器通信连接的存储器;其中,
    所述存储器存储有可被所述至少一个处理器执行的智能回访程序,所述智能回访程序被所述至少一个处理器执行,以使所述至少一个处理器能够执行如下步骤:
    解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
    基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
    获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
    基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
  10. 如权利要求9所述的电子设备,其中,所述基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷包括:
    从所述用户数据中抽取预设指标项对应的数据生成第一问卷;
    从第二数据库中获取所述回访类别对应的第二问卷,基于所述用户画像从所述第二问卷中抽取第三问卷;
    合并所述第一问卷及第三问卷,得到所述待回访用户对应的回访问卷。
  11. 如权利要求9所述的电子设备,其中,所述基于所述第一特征确定所述待回访用户对应的目标回访渠道包括:
    对所述第一特征执行降维处理,得到第二特征;
    获取第三数据库中已执行聚类分析的各个用户组的中心用户的第三特征;
    基于所述第二特征及第三特征确定所述待回访用户对应的目标用户组;
    基于所述目标用户组及用户组与回访渠道对应的映射关系确定所述待回访用户对应的目标回访渠道。
  12. 如权利要求11所述的电子设备,其中,所述聚类分析包括:
    获取第三数据库中各个用户的历史数据,基于所述特征因子及所述历史数据确定各个用户的第四特征;
    基于所述第四特征及K均值聚类算法对所述第三数据库中的用户进行分组,其中,K分别取值为预设数值范围内的各个自然数,K的一个取值对应一种分组结果,得到多种分组结果;
    基于所述多种分组结果中每种分组结果对应的各个用户组的中心用户的第四特征计算每种分组结果对应的轮廓系数;
    将轮廓系数最接近预设数值的分组结果作为目标分组结果。
  13. 如权利要求9所述的电子设备,其中,所述目标回访渠道包括AI语音回访及网页回访,所述基于所述目标回访渠道对所述待回访用户进行身份验证包括:
    若所述回访渠道为AI语音回访,连接所述标识对应的第二客户端,获取所述第二客户端的用户在第一预设时间段内的第一音频数据,基于所述第一音频数据对所述待回访用户进行身份验证;
    若所述回访渠道为网页回访,连接所述标识对应的第三客户端,获取所述第三客户端的用户在第二预设时间段内的视频数据,基于所述视频数据对所述待回访用户进行身份验证。
  14. 如权利要求13所述的电子设备,其中,所述基于所述第一音频数据对所述待回访用户进行身份验证包括:
    对所述第一音频数据进行短时傅里叶变换和/或短时傅里叶逆变换,得到所述第二客户端的用户的时域信号数据,将所述时域信号数据输入声纹识别模型,得到目标声纹特征,从第四数据库中获取所述待回访用户对应的标准声纹特征,若所述目标声纹特征与标准声纹特征的相似度值小于声纹相似度阈值,则判断所述待回访用户身份验证通过;
    所述基于所述视频数据对所述待回访用户进行身份验证包括:
    对所述视频数据进行分帧,得到图像序列,将所述图像序列输入人脸识别模型,得到目标人脸特征,从第五数据库中获取所述待回访用户对应的标准人脸特征,若所述目标人脸特征与标准人脸特征的相似度值小于人脸相似度阈值,则判断所述待回访用户身份验证通过。
  15. 如权利要求12所述的电子设备,其中,所述轮廓系数对应的计算公式为:
    Figure PCTCN2021123905-appb-100005
    Figure PCTCN2021123905-appb-100006
    其中,S ij表示第i种分组结果中第j个用户对应的轮廓系数,
    Figure PCTCN2021123905-appb-100007
    表示第i种分组结果中第j个用户的第四特征到同一个用户组中其他用户的第四特征的平均距离,
    Figure PCTCN2021123905-appb-100008
    表示第i种分组结果中第j个用户的第四特征到其他用户组的第四特征的平均距离的最小值,S i表示第i种分组结果对应的轮廓系数,n表示用户的总数量。
  16. 一种计算机可读存储介质,其中,所述计算机可读存储介质上存储有智能回访程序,所述智能回访程序可被一个或者多个处理器执行,以实现如下步骤:
    解析第一客户端发出的回访请求,获取所述回访请求携带的待回访用户的标识,基于所述标识从第一数据库中获取所述待回访用户的用户数据;
    基于所述用户数据确定所述待回访用户对应的回访类别,并为所述待回访用户建立用户画像,基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷;
    获取所述回访类别对应的特征因子,根据所述特征因子得到所述待回访用户的第一特征,基于所述第一特征确定所述待回访用户对应的目标回访渠道;
    基于所述目标回访渠道对所述待回访用户进行身份验证,当身份验证通过时,基于所述目标回访渠道及所述回访问卷对所述待回访用户进行回访,并生成回访报告。
  17. 如权利要求16所述的计算机可读存储介质,其中,所述基于所述用户画像及所述回访类别为所述待回访用户生成回访问卷包括:
    从所述用户数据中抽取预设指标项对应的数据生成第一问卷;
    从第二数据库中获取所述回访类别对应的第二问卷,基于所述用户画像从所述第二问卷中抽取第三问卷;
    合并所述第一问卷及第三问卷,得到所述待回访用户对应的回访问卷。
  18. 如权利要求16所述的计算机可读存储介质,其中,所述基于所述第一特征确定所述待回访用户对应的目标回访渠道包括:
    对所述第一特征执行降维处理,得到第二特征;
    获取第三数据库中已执行聚类分析的各个用户组的中心用户的第三特征;
    基于所述第二特征及第三特征确定所述待回访用户对应的目标用户组;
    基于所述目标用户组及用户组与回访渠道对应的映射关系确定所述待回访用户对应的目标回访渠道。
  19. 如权利要求18所述的计算机可读存储介质,其中,所述聚类分析包括:
    获取第三数据库中各个用户的历史数据,基于所述特征因子及所述历史数据确定各个用户的第四特征;
    基于所述第四特征及K均值聚类算法对所述第三数据库中的用户进行分组,其中,K分别取值为预设数值范围内的各个自然数,K的一个取值对应一种分组结果,得到多种分组结果;
    基于所述多种分组结果中每种分组结果对应的各个用户组的中心用户的第四特征计算每种分组结果对应的轮廓系数;
    将轮廓系数最接近预设数值的分组结果作为目标分组结果。
  20. 如权利要求16所述的计算机可读存储介质,其中,所述目标回访渠道包括AI语音回访及网页回访,所述基于所述目标回访渠道对所述待回访用户进行身份验证包括:
    若所述回访渠道为AI语音回访,连接所述标识对应的第二客户端,获取所述第二客户端的用户在第一预设时间段内的第一音频数据,基于所述第一音频数据对所述待回访用户进行身份验证;
    若所述回访渠道为网页回访,连接所述标识对应的第三客户端,获取所述第三客户端的用户在第二预设时间段内的视频数据,基于所述视频数据对所述待回访用户进行身份验证。
PCT/CN2021/123905 2020-11-19 2021-10-14 智能回访方法、装置、电子设备及可读存储介质 WO2022105496A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011305616.1 2020-11-19
CN202011305616.1A CN112507212A (zh) 2020-11-19 2020-11-19 智能回访方法、装置、电子设备及可读存储介质

Publications (1)

Publication Number Publication Date
WO2022105496A1 true WO2022105496A1 (zh) 2022-05-27

Family

ID=74959933

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/123905 WO2022105496A1 (zh) 2020-11-19 2021-10-14 智能回访方法、装置、电子设备及可读存储介质

Country Status (2)

Country Link
CN (1) CN112507212A (zh)
WO (1) WO2022105496A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114881712A (zh) * 2022-07-12 2022-08-09 杭州新航互动科技有限公司 智能广告投放方法、装置、设备及存储介质

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111080312A (zh) * 2019-11-15 2020-04-28 泰康保险集团股份有限公司 用户回访方法、装置、电子设备及存储介质
CN112507212A (zh) * 2020-11-19 2021-03-16 深圳壹账通智能科技有限公司 智能回访方法、装置、电子设备及可读存储介质
CN114970552B (zh) * 2022-07-27 2022-10-11 成都乐超人科技有限公司 基于微服务的用户回访信息分析方法、装置、设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9002839B1 (en) * 2007-09-28 2015-04-07 Amazon Technologies, Inc. Personalizing content for users
CN105376432A (zh) * 2015-09-24 2016-03-02 深圳市车音网科技有限公司 语音自动回访装置、系统及其方法
CN107682658A (zh) * 2017-09-15 2018-02-09 平安科技(深圳)有限公司 一种视频回访方法、设备及存储介质
CN109285011A (zh) * 2018-09-12 2019-01-29 田雪娇 一种智能客户回访方法、装置及设备
CN112507212A (zh) * 2020-11-19 2021-03-16 深圳壹账通智能科技有限公司 智能回访方法、装置、电子设备及可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9002839B1 (en) * 2007-09-28 2015-04-07 Amazon Technologies, Inc. Personalizing content for users
CN105376432A (zh) * 2015-09-24 2016-03-02 深圳市车音网科技有限公司 语音自动回访装置、系统及其方法
CN107682658A (zh) * 2017-09-15 2018-02-09 平安科技(深圳)有限公司 一种视频回访方法、设备及存储介质
CN109285011A (zh) * 2018-09-12 2019-01-29 田雪娇 一种智能客户回访方法、装置及设备
CN112507212A (zh) * 2020-11-19 2021-03-16 深圳壹账通智能科技有限公司 智能回访方法、装置、电子设备及可读存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114881712A (zh) * 2022-07-12 2022-08-09 杭州新航互动科技有限公司 智能广告投放方法、装置、设备及存储介质
CN114881712B (zh) * 2022-07-12 2022-09-23 杭州新航互动科技有限公司 智能广告投放方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN112507212A (zh) 2021-03-16

Similar Documents

Publication Publication Date Title
WO2022105496A1 (zh) 智能回访方法、装置、电子设备及可读存储介质
WO2022105115A1 (zh) 问答对匹配方法、装置、电子设备及存储介质
US20230013306A1 (en) Sensitive Data Classification
WO2021068610A1 (zh) 资源推荐的方法、装置、电子设备及存储介质
WO2021174944A1 (zh) 基于目标对象活跃度的消息推送方法及相关设备
US11461298B1 (en) Scoring parameter generation for identity resolution
WO2022078308A1 (zh) 裁判文书摘要生成方法、装置、电子设备及可读存储介质
CN112287244A (zh) 基于联邦学习的产品推荐方法、装置、计算机设备及介质
WO2022089190A1 (zh) 产品推荐方法、装置、电子设备及可读存储介质
WO2018201601A1 (zh) 基于数据源的业务定制装置、方法、系统及存储介质
CN110929525B (zh) 一种网贷风险行为分析检测方法、装置、设备和存储介质
CN112528315B (zh) 识别敏感数据的方法和装置
CN112364937A (zh) 用户类别确定方法及装置、推荐内容确定方法、电子设备
CN114387061A (zh) 产品推送方法、装置、电子设备及可读存储介质
CN112214515B (zh) 数据自动匹配方法、装置、电子设备及存储介质
CN111339784B (zh) 一种新话题的自动挖掘方法和系统
CN111552865A (zh) 用户兴趣画像方法及相关设备
US10902215B1 (en) Social hash for language models
JP2023517518A (ja) ヌル値又は同等の値を有するリレーショナル・テーブルのためのベクトル埋込モデル
CN111222032A (zh) 舆情分析方法及相关设备
CN116844711A (zh) 基于深度学习的疾病辅助识别方法及装置
CN111597453B (zh) 用户画像方法、装置、计算机设备及计算机可读存储介质
CN112084408B (zh) 名单数据筛选方法、装置、计算机设备及存储介质
CN114254112A (zh) 用于敏感信息预分类的方法、系统、装置和介质
US11989243B2 (en) Ranking similar users based on values and personal journeys

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: 21893645

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 22.08.2023)

122 Ep: pct application non-entry in european phase

Ref document number: 21893645

Country of ref document: EP

Kind code of ref document: A1