Call processing method and apparatus, electronic device, and storage medium

Abstract

Embodiments of the present application provide a call processing method and device, electronic equipment and storage medium, relating to the field of financial technology or other related fields, the method comprising: recording a call recording in the process of a user's incoming call complaint; obtaining a keyword related to the user's complaint purpose and an emotional feature of the user according to the call recording, the emotional feature being used to indicate the emotional tendency of the user; obtaining at least one complaint processing scheme from a preset scheme library according to the keyword; and outputting the emotional feature and the at least one complaint processing scheme to a terminal device corresponding to a customer service, so as to be referred to in the process of communication between the customer service and the user. Through the above method, the probability that the customer service cannot provide a satisfactory solution for the user due to the deviation of the emotional judgment of the customer service for the user and / or the low familiarity with the business processing flow can be reduced, thereby providing high-quality service for the user.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of financial technology or other related fields, and in particular, to a call processing method and device, an electronic device, and a storage medium. BACKGROUND

[0002] A customer service center based on telephone communication is an important way for enterprises to communicate with users. Handling user telephone complaints is an important responsibility of the customer service center. In order to provide a high-quality complaint handling solution for users, not only does the customer service personnel need to be familiar with various business process specifications, but also needs to have a keen judgment ability for the user's emotions.

[0003] For example, the user's satisfaction solution may be different when the user is angry and when the user is angry. If the customer service personnel deviates in judging the user's emotions according to their own experience, and / or, is not familiar with the business process, the solution provided may not meet the user's needs, so that the user thinks that the service quality of the customer service is poor, and causes complaint escalation. SUMMARY

[0004] The embodiments of the present application provide a call processing method, device, electronic device, and storage medium, which can reduce the deviation of the customer service personnel in judging the user's emotions when handling user complaints, and provide a reference processing solution to the customer service personnel, thereby improving the service quality of the customer service.

[0005] In a first aspect, the embodiments of the present application provide a call processing method, comprising:

[0006] recording a call recording in the process of a user's incoming call complaint;

[0007] obtaining a keyword related to the user's complaint purpose and an emotional feature of the user according to the call recording, the emotional feature being used to indicate the user's emotional tendency;

[0008] obtaining at least one complaint handling solution from a preset solution library according to the keyword;

[0009] outputting the emotional feature and at least one complaint handling solution to a terminal device corresponding to the customer service, so as to be referred to in the process of communication between the customer service and the user.

[0010] Optionally, the obtaining a keyword related to the user's complaint purpose and an emotional feature of the user according to the call recording comprises:

[0011] obtaining a voiceprint feature included in the call recording;

[0012] According to the voiceprint feature and a preset voiceprint feature of the customer service, determine voice data of a user in the call recording;

[0013] According to the voice data of the user, determine the keyword and the emotional feature of the user.

[0014] Optionally, the determining the emotional feature of the user according to the voice data of the user comprises:

[0015] performing semantic analysis on the voice data of the user to obtain a sentiment tendency distribution probability of the user;

[0016] taking a sentiment tendency corresponding to a preset threshold in the sentiment tendency distribution probability as the emotional feature of the user.

[0017] Optionally, the performing semantic analysis on the voice data of the user to obtain the sentiment tendency distribution probability of the user comprises:

[0018] inputting the voice data of the user into an emotion classification model to obtain an initial sentiment tendency distribution probability of the user;

[0019] inputting the voice data of the user into an emotion recognition model to obtain an emotion curve of the user;

[0020] correcting the initial sentiment tendency distribution probability according to the emotion curve to obtain the sentiment tendency distribution probability.

[0021] Optionally, the emotion classification model comprises a first sub-model, a second sub-model and a classification function.

[0022] The first sub-model is configured to obtain semantic information of the voice data of the user according to the voice data of the user.

[0023] The second sub-model is configured to obtain a semantic feature vector according to the semantic information.

[0024] The classification function is configured to map the semantic feature vector to obtain the initial sentiment tendency distribution probability of the user.

[0025] Optionally, the first sub-model comprises a processing unit and a modeling unit.

[0026] The processing unit is configured to convert the voice data of the user into text data, and perform word segmentation processing on the text data to obtain an intention text of the user.

[0027] The modeling unit is configured to use a Trm encoder to model relationships of words in the intention text to obtain the semantic information.

[0028] Optionally, the second sub-model comprises: a bidirectional simple recurrent unit (BiSRU) and a soft attention unit.

[0029] The BiSRU is configured to obtain an initial semantic feature vector according to the semantic information.

[0030] The soft attention unit is configured to obtain the semantic feature vector according to the initial semantic feature vector.

[0031] In a second aspect, an embodiment of the present application provides a call processing apparatus, comprising:

[0032] A recording module is configured to record a call recording in a process in which a user calls to complain.

[0033] A first obtaining module is configured to obtain a keyword related to a complaint purpose of the user and an emotional feature of the user according to the call recording, the emotional feature being used to indicate an emotional tendency of the user.

[0034] A second obtaining module is configured to obtain at least one complaint processing scheme from a preset scheme library according to the keyword.

[0035] An output module is configured to output the emotional feature and the at least one complaint processing scheme to a terminal device corresponding to a customer service, so that the customer service can refer to the emotional feature and the at least one complaint processing scheme in a process of communicating with the user.

[0036] Optionally, the call processing apparatus can execute the call processing method of any one of the first aspect.

[0037] In a third aspect, the present application provides an electronic device, comprising: a memory and a processor.

[0038] The memory is configured to store computer instructions, and the processor is configured to execute the computer instructions stored in the memory to implement the method of any one of the first aspect.

[0039] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method of any one of the first aspect.

[0040] In a fifth aspect, the present application provides a computer program product, comprising a computer program, and the computer program is executed by a processor to implement the method of any one of the first aspect.

[0041] The embodiment of the present application provides a call processing method and device, electronic equipment and storage medium. By analyzing the voice of a user, the current emotional characteristics of the user and at least one complaint processing scheme can be output to a customer service, so that the probability that the user thinks that the service quality of the customer service is poor is reduced due to the deviation of the emotional judgment of the customer service to the user and / or the low familiarity with the business processing flow. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 Flowchart of the call processing method provided by the embodiment of the present application Figure 1 ;

[0043] Figure 2 Flowchart of the call processing method provided by the embodiment of the present application Figure 2 ;

[0044] Figure 3 Structure diagram of the emotion classification model provided by the embodiment of the present application;

[0045] Figure 4 Structure diagram of the call processing device provided by the embodiment of the present application;

[0046] Figure 5 Structure diagram of the electronic equipment provided by the embodiment of the present application. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0048] In the embodiments of the present application, the same items or similar items with basically the same functions and effects are distinguished by using "first", "second", etc., and the sequence is not limited. Those skilled in the art can understand that "first", "second", etc. do not limit the quantity and execution sequence, and "first", "second", etc. do not necessarily mean different.

[0049] It should be noted that in the embodiments of the present application, "exemplary" or "for example" is used to represent an example, illustration or description. Any embodiment or design scheme described as "exemplary" or "for example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, "exemplary" or "for example" is used to present the relevant concept in a specific manner.

[0050] The customer service center based on telephone communication is an important way for enterprises to communicate with users. Handling telephone complaints of users is an important responsibility of the customer service center. The satisfaction degree of users to the handling result of complaints can reflect the degree of attention of the enterprise to the users to some extent.

[0051] In order to provide a high-quality complaint handling scheme for users, not only the customer service personnel should be familiar with various business process specifications, but also the customer service personnel should have sharp judgment ability on the emotions of users. For example, the solution that satisfies the user can be different when the user is angry and when the user is furious.

[0052] At present, the handling of user complaints basically depends on the experience of customer service personnel in handling complaints. If the customer service personnel deviates in judging the emotions of users according to their own experience, and / or is not familiar with the business process, the solution given can not meet the needs of users, so that the users think that the service quality of the customer service is poor, and the complaint escalation is triggered.

[0053] Therefore, the embodiment of the present application provides a call processing method and device, electronic equipment and storage medium. By recording the call process of the customer service and the user, using a preset voice processing model to analyze the user voice in the recording, the current emotional state of the user and at least one complaint handling scheme can be obtained, and the current emotional state of the user and at least one complaint handling scheme are output to the customer service, so as to reduce the probability that the user thinks that the service quality of the customer service is poor due to the deviation of the customer service in judging the emotions of the user and / or the customer service is not familiar with the business process.

[0054] The technical solutions of the present application and how the technical solutions of the present application solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be implemented independently, or can be combined with each other. For the same or similar concepts or processes, some embodiments can not be described again.

[0055] Figure 1 The flowchart of the call processing method provided by the embodiment of the present application is shown as follows. Figure 1 As shown in the figure, the following steps are included:

[0056] S101, record the call recording in the process of user incoming call complaint.

[0057] The execution subject of the embodiment of the present application can be a voice processing system. The recording of the call recording can be performed according to the preset time, for example, recording a call recording every one minute. The voice processing system can analyze each call recording.

[0058] S102, obtain a keyword related to the user complaint purpose and an emotion feature of the user according to the call recording, and the emotion feature is used to indicate an emotional tendency of the user.

[0059] In the embodiment of the application, the keyword related to the user complaint purpose can be used to indicate the user complaint content. The emotion feature of the user can be a feature indicating the emotional tendency of the user, for example, anger, anger, etc.

[0060] In the embodiment of the application, the voice processing system can obtain the keyword related to the user complaint purpose and the emotion feature of the user according to the voice of the user in the call recording.

[0061] For example, the voice processing system can use speech recognition technology to identify the content related to the complaint event name included in the voice of the user to determine the keyword related to the complaint purpose. For example, the voice of the user includes "what is the situation of your business hall, why do you need so many data to handle a bank card", and the voice processing system can determine that the keyword related to the complaint purpose is "too much data to handle a bank card".

[0062] The voice processing system can input the voice of the user into a pre-trained emotion processing model to obtain the emotion feature of the user output by the emotion processing model. For example, the emotion processing model can include Ro-BERTa model and BiSRU-AT model.

[0063] S103, obtain at least one complaint processing scheme from a pre-set scheme library according to the keyword.

[0064] In the embodiment of the application, the voice processing system can determine at least one complaint processing scheme including the keyword from the scheme library according to the determined keyword. For example, by keyword matching each complaint processing scheme in the pre-set scheme library, at least one complaint processing scheme including the keyword is determined.

[0065] S104, output the emotion feature and at least one complaint processing scheme to the terminal device corresponding to the customer service, so as to be referred to in the process of communication between the customer service and the user.

[0066] In the embodiment of the application, when the voice processing system determines the emotion feature of the user and at least one complaint processing scheme, the emotion feature and at least one complaint processing scheme can be pushed to the terminal device of the customer service personnel, so that the customer service personnel can refer to the current emotion feature of the user and the complaint processing scheme provided by the system to provide appropriate solutions to the customer.

[0067] The call processing method provided in the embodiments of the present application can record a call recording in the process of a user's incoming call complaint, acquire a keyword related to the user's complaint purpose and an emotional feature of the user according to the call recording, the emotional feature being used to indicate the emotional tendency of the user, acquire at least one complaint processing scheme from a preset scheme library according to the keyword, and output the emotional feature and the at least one complaint processing scheme to a terminal device corresponding to a customer service, so as to be referred to in the process of communication between the customer service and the user. Through the above method, the probability that the user thinks that the service quality of the customer service is poor can be reduced due to the deviation of the emotional judgment of the customer service on the user and / or the low familiarity with the business processing flow, which cannot provide a satisfactory solution for the user.

[0068] Figure 2 Flowchart of the call processing method provided in the embodiments of the present application Figure 2 In the process of Figure 1 On the basis of the embodiments shown in the above, the process of determining the keyword related to the user's complaint purpose and the emotional feature of the user is further described, as shown in the following. Figure 2 The process includes the following steps:

[0069] S201, acquire the voice data of the user in the call recording.

[0070] In the embodiments of the present application, different people have different voiceprint features due to the particularity of sound, wherein the voiceprint feature can be a voice feature that can represent and identify a speaker. The voice data of the user in the call recording can be acquired through the voiceprint feature.

[0071] For example, the voiceprint feature included in the call recording is acquired, and the voice data of the user in the call recording is determined according to the voiceprint feature and a preset voiceprint feature of the customer service.

[0072] The voice processing system can obtain the voiceprint features of the customer service and the user according to the voiceprint recognition of the call recording, distinguish the customer service and the user in the call recording through the preset voiceprint feature of the customer service, and thus obtain the voice data of the user in the call recording.

[0073] S202, acquire the keyword related to the complaint purpose according to the voice data of the user, and acquire at least one complaint processing scheme from a preset scheme library according to the keyword.

[0074] In the embodiments of the present application, when the voice processing system acquires the voice data of the user, the keyword related to the complaint purpose can be determined according to the name of the related business mentioned in the voice data of the user.

[0075] Exemplarily, the user voice data is converted into corresponding text content by using a voice recognition technology; and the text content is matched according to a preset business name to determine the keyword.

[0076] In the embodiment of the application, the keyword can indicate a business or matter complained by the user. When the voice processing system obtains the keyword, the voice processing system can determine an identifier corresponding to the business or matter indicated by the keyword, and find a complaint processing scheme matching the identifier from a preset scheme library according to the identifier, and filter out all complaint processing schemes including the identifier, so as to determine at least one complaint processing scheme.

[0077] S203, performing semantic analysis on the voice data of the user to obtain a probability distribution of emotional tendency of the user.

[0078] In the embodiment of the application, the emotional tendency can indicate which state the user is currently in, and the probability distribution of emotional tendency of the user can be a probability distribution of different emotional tendencies of the current user determined according to the voice data of the user. For example, after performing semantic analysis on the voice data of the user, it can be determined that the probability of the user being currently in an angry state is 70%, the probability of the user being currently in an angry state is 25%, and the probability of the user being currently in a calm state is 5%.

[0079] The voice data of the user can be analyzed semantically according to the following manner to obtain the probability distribution of emotional tendency of the user.

[0080] Exemplarily, the voice data of the user is input into an emotion classification model to obtain an initial probability distribution of emotional tendency of the user; the voice data of the user is input into an emotion recognition model to obtain an emotion curve of the user; and the initial probability distribution of emotional tendency is corrected according to the emotion curve to obtain the probability distribution of emotional tendency.

[0081] In the embodiment of the application, the emotion classification model can be a pre-trained emotion tendency judgment model, for example, a BERT model based on Transformer. The voice data of the user can be analyzed semantically by the emotion classification model to determine the probability distribution of emotional tendency of the text content corresponding to the voice data of the user. The emotion recognition model can be a model for analyzing changes in user emotions based on a neural network model, for example, a back propagation BP neural network model. The emotion curve can be a curve reflecting changes in user emotions.

[0082] In the embodiment of the application, the user's speech content can be used to determine the user's current emotions to some extent, and the user's tone features can also be used to determine the user's current emotions to some extent.

[0083] The voice processing system can input the voice data of the user into an emotion classification model and an emotion recognition model respectively to obtain an initial emotion tendency distribution probability of the user output by the emotion classification model and an emotion curve of the user output by the emotion recognition model. The initial emotion tendency distribution probability is corrected through the emotion curve of the user to obtain a more accurate emotion tendency distribution probability of the user.

[0084] For example, the initial emotion tendency distribution probability of the user is that the probability of anger is 70%, the probability of anger is 25%, and the probability of peace is 5%. According to the emotion curve of the user, the proportion of different emotions of the user is determined as follows: the proportion of anger is 65%, the proportion of peace is 10%, the proportion of happiness is 5%, and the proportion of anger is 20%. The initial emotion tendency distribution probability and the emotion of each user in the emotion curve are analyzed comprehensively by using the analytic hierarchy process to obtain the emotion tendency distribution probability of the user. For example, the final emotion tendency distribution probability of the user can be that the probability of anger is 68%, the probability of anger is 22%, the probability of peace is 8%, and the probability of happiness is 2%.

[0085] The specific structure of the emotion classification model will be introduced below.

[0086] In the embodiment of the application, the emotion classification model includes a first sub-model, a second sub-model and a classification function. The first sub-model is configured to obtain semantic information of the voice data of the user according to the voice data of the user. The second sub-model is configured to obtain a semantic feature vector according to the semantic information. The classification function is configured to map the semantic feature vector to obtain an initial emotion tendency distribution probability of the user.

[0087] For example, the structure of the emotion classification model is as shown in Figure 3 The first sub-model can be a Ro-BERTa model, the second sub-model can be a BiSRU-AT model, and the classification function can be a softmax classification function.

[0088] The first sub-model includes a processing unit and a modeling unit. The processing unit is configured to convert the voice data of the user into text data and perform word segmentation processing on the text data to obtain an intent text of the user. The modeling unit is configured to use a Trm encoder to model the relationship of each word in the intent text to obtain the semantic information.

[0089] The second sub-model includes a bidirectional simple recurrent unit BiSRU and a soft attention unit. The BiSRU is configured to obtain an initial semantic feature vector according to the semantic information. The soft attention unit is configured to obtain the semantic feature vector according to the initial semantic feature vector.

[0090] For example, please continue to refer toFigure 3 The voice processing system can input the voice data of the user as an input of the emotion classification model, and perform text conversion on the voice data of the user by using a processing unit in the Ro-BERTa model. The converted text is subjected to word-level segmentation by using a word segmenter, and words are converted into corresponding serial numbers according to a vocabulary table to obtain an intent text of the user. A modeling unit in the Ro-BERTa model uses a Transformer encoder to model the relationship between words of any length in the intent text, capture the grammatical structure inside the sentence and the dependency relationship between words, and obtain corresponding semantic information. The semantic information can be used as an input of a second sub-model.

[0091] The BiSRU is composed of forward and reverse simple recurrent units (SRUs), and the outputs of the forward and reverse SRUs are combined to obtain the output of the BiSRU. The semantic information is input into the BiSRU to obtain a row vector of the output, that is, an initial semantic feature vector. The initial semantic feature vector is input into a soft attention unit to obtain a corresponding semantic feature vector. Finally, the semantic feature vector is mapped by using a softmax classification function to obtain a probability of emotion tendency distribution of the user.

[0092] S204, determining an emotional feature of the user according to the probability of emotion tendency distribution of the user.

[0093] In the embodiment of the present application, when the voice processing system obtains the probability of emotion tendency distribution of the model output, the emotion tendency corresponding to a probability greater than or equal to a preset threshold in the probability of emotion tendency distribution can be used as the emotional feature of the user. The preset threshold can be set according to actual needs, and the embodiment of the present application does not limit this.

[0094] S205, outputting the emotional feature and at least one complaint processing scheme to a terminal device corresponding to the customer service.

[0095] The specific implementation mode shown in S205 in the embodiment of the present application is similar to the specific implementation mode of S104 in the embodiment shown in Figure 1 The specific implementation mode shown in S205 in the embodiment of the present application is similar to the specific implementation mode of S104 in the embodiment shown in

[0096] The call processing method provided by the embodiments of the present application can obtain the emotional characteristics of the user by comprehensively processing the voice data of the user in the call recording through the emotion classification model and the emotion recognition model, improve the accuracy of the positioning of the user's emotion, obtain at least one complaint processing scheme from the preset scheme library according to the keywords related to the complaint purpose, and improve the probability. The emotional characteristics of the user and the at least one complaint processing scheme are output to the terminal device of the customer service. Thus, the probability that the customer service deviates in judging the emotion of the user and / or is not familiar with the business processing flow and cannot provide a satisfactory solution for the user is reduced, thereby providing high-quality service for the user.

[0097] The embodiments of the present application also provide a call processing device.

[0098] Figure 4 The structural schematic diagram of the call processing device 40 provided by the embodiments of the present application is shown in Figure 4 , and includes:

[0099] The recording module 401 is configured to record a call recording in the process of the user calling to complain.

[0100] The first obtaining module 402 is configured to obtain a keyword related to the complaint purpose of the user and an emotional characteristic of the user according to the call recording, and the emotional characteristic is used to indicate the emotional tendency of the user.

[0101] The second obtaining module 403 is configured to obtain at least one complaint processing scheme from a preset scheme library according to the keyword.

[0102] The output module 404 is configured to output the emotional characteristic and the at least one complaint processing scheme to the terminal device corresponding to the customer service for reference in the process of the customer service communicating with the user.

[0103] Optionally, the first obtaining module 402 is further configured to obtain a voiceprint characteristic included in the call recording, determine the voice data of the user in the call recording according to the voiceprint characteristic and a preset voiceprint characteristic of the customer service, and determine the keyword and the emotional characteristic of the user according to the voice data of the user.

[0104] Optionally, the first obtaining module 402 is further configured to perform semantic analysis on the voice data of the user to obtain a probability of the emotional tendency distribution of the user, and take an emotional tendency corresponding to a preset threshold in the probability of the emotional tendency distribution as the emotional characteristic of the user.

[0105] Optionally, the first obtaining module 402 is further configured to input the voice data of the user into an emotion classification model to obtain an initial emotion tendency distribution probability of the user; input the voice data of the user into an emotion recognition model to obtain an emotion curve of the user; and correct the initial emotion tendency distribution probability according to the emotion curve to obtain the emotion tendency distribution probability.

[0106] Optionally, the first obtaining module 402 is further configured to obtain semantic information of the voice data of the user according to the voice data of the user; obtain a semantic feature vector according to the semantic information; and map the semantic feature vector to obtain the initial emotion tendency distribution probability of the user.

[0107] Optionally, the first obtaining module 402 is further configured to convert the voice data of the user into text data, and perform word segmentation processing on the text data to obtain an intention text of the user; and model relationships of words in the intention text by using a Trm encoder to obtain the semantic information.

[0108] Optionally, the first obtaining module 402 is further configured to obtain an initial semantic feature vector according to the semantic information; and obtain the semantic feature vector according to the initial semantic feature vector.

[0109] The call processing apparatus provided by the embodiments of the present application can execute the technical solutions of the call processing method provided by any of the above embodiments, and has similar principles and technical effects, which will not be described herein again.

[0110] Figure 5 An electronic device structure schematic diagram is provided in the present application. As shown in the figure, the electronic device 50 can include at least one processor 501, a memory 502. Figure 5

[0111] The memory 502 is used to store programs. Specifically, the programs can include program codes, and the program codes include computer operation instructions.

[0112] The memory 502 can include a high-speed RAM memory, and can also include a non-volatile memory, for example, at least one disk memory.

[0113] ​The processor 501 is configured to execute the computer-executable instructions stored in the memory 502, so as to implement the call processing method described in the foregoing method embodiments. The processor 501 can be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0114] Optionally, the electronic device 50 further includes a communication interface 503. In a specific implementation, if the communication interface 503, the memory 502, and the processor 501 are independently implemented, the communication interface 503, the memory 502, and the processor 501 can be connected with each other through a bus and complete communication therebetween. The bus can be an industry standard architecture (ISA) bus, a peripheral component (PCI) bus, or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc., but does not mean that there is only one bus or only one type of bus.

[0115] Optionally, in a specific implementation, if the communication interface 503, the memory 502, and the processor 501 are integrated on a chip, the communication interface 503, the memory 502, and the processor 501 can complete communication through an internal interface.

[0116] The embodiments of the present application further provide a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the technical solutions of the call processing method embodiments, and the implementation principles and technical effects are similar, which will not be described herein.

[0117] In a possible implementation, the computer readable medium can include a random access memory (RAM), a read-only memory (ROM), a compact disc read-only memory (CD-ROM) or other optical memory, a magnetic disk storage or other magnetic storage devices, or any other medium that is suitable for storing desired program code in the form of instructions or data structures and that can be accessed by a computer. Moreover, any connection is properly termed a computer readable medium. For example, if the software is transmitted from a website, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. Disk and disc, as used herein, include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray® disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer readable media.

[0118] The embodiments of the present application also provide a computer program product, including a computer program, which, when executed by a processor, implements the technical solutions of the call processing method embodiments described above, and has similar implementation principles and technical effects, which will not be described here.

[0119] In the specific implementation of the terminal device or the server described above, it should be understood that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as execution completed by a hardware processor, or executed by a combination of hardware and software modules in the processor.

[0120] Those skilled in the art can understand that all or part of the steps of any of the method embodiments described above can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer readable storage medium, and when the program is executed, all or part of the steps of the method embodiments described above are executed.

[0121] If the technical solutions of the present application are implemented in the form of software and sold or used as products, they can be stored in a computer readable storage medium. Based on this understanding, all or part of the technical solutions of the present application can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a computer program or a plurality of instructions. The computer software product enables a computer device (which can be a personal computer, a server, a network device or similar electronic equipment) to execute all or part of the steps of the method described in the embodiments of the present application.

[0122] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A call processing method characterized by, The method comprises the following steps: recording a call recording in the process of a user calling to complain; obtaining a voiceprint feature included in the call recording; determining voice data of the user in the call recording according to the voiceprint feature and a preset voiceprint feature of a customer service; obtaining a keyword related to a complaint purpose of the user and an emotion feature of the user according to the voice data of the user in the call recording, the emotion feature being used to indicate an emotional tendency of the user; obtaining at least one complaint processing scheme from a preset scheme library according to the keyword; outputting the emotion feature and the at least one complaint processing scheme to a terminal device corresponding to the customer service, so as to be referred to in the process of the customer service communicating with the user; obtaining the emotion feature of the user according to the voice data of the user in the call recording, comprising: inputting the voice data into an emotion classification model to obtain an initial emotional tendency distribution probability of the user; the emotion classification model comprises a first sub-model, a second sub-model and a classification function; the first sub-model is used to obtain semantic information of the voice data of the user according to the voice data of the user; the second sub-model is used to obtain a semantic feature vector according to the semantic information; and the classification function is used to map the semantic feature vector to obtain the initial emotional tendency distribution probability of the user; inputting the voice data into an emotion recognition model to obtain an emotion curve of the user; the emotion recognition model is a back propagation (BP) neural network model; and the emotion curve is used to reflect the change of the emotion of the user; correcting the initial emotional tendency distribution probability according to the emotion curve by using an analytic hierarchy process to obtain an emotional tendency distribution probability; taking an emotion tendency corresponding to a preset threshold in the emotional tendency distribution probability as the emotion feature of the user.

2. The method of claim 1, wherein, The first sub-model comprises a processing unit and a modeling unit; the processing unit is used to convert the voice data of the user into text data, and perform word segmentation processing on the text data to obtain an intention text of the user; the modeling unit is used to model the relationship of each word in the intention text by using a Trm encoder to obtain the semantic information.

3. The method of claim 1, wherein, The second sub-model comprises a bidirectional simple recurrent unit (BiSRU) and a soft attention unit; the BiSRU is used to obtain an initial semantic feature vector according to the semantic information; the soft attention unit is used to obtain the semantic feature vector according to the initial semantic feature vector.

4. A call processing apparatus characterized by comprising: The method comprises the following steps: a recording module is configured to record a call recording in the process of a user calling to complain; a first obtaining module is configured to obtain a keyword related to a complaint purpose of the user and an emotion feature of the user according to voice data of the user in the call recording, the emotion feature being used to indicate an emotional tendency of the user; a second obtaining module is configured to obtain at least one complaint processing scheme from a preset scheme library according to the keyword. An output module is configured to output the emotional feature and at least one complaint processing scheme to a terminal device corresponding to the customer service, so as to be referred to in the process of communication between the customer service and the user. The first obtaining module is specifically configured to input the voice data into an emotion classification model to obtain an initial emotional tendency distribution probability of the user. The emotion classification model comprises a first sub-model, a second sub-model and a classification function. The first sub-model is configured to obtain semantic information of the voice data of the user according to the voice data of the user; the second sub-model is configured to obtain a semantic feature vector according to the semantic information; and the classification function is configured to map the semantic feature vector to obtain the initial emotional tendency distribution probability of the user. The voice data is input into an emotion recognition model to obtain an emotional curve of the user; the emotion recognition model is a back propagation (BP) neural network model; the emotional curve is used to reflect the change of the user's emotion; the initial emotional tendency distribution probability is corrected by using an analytic hierarchy process according to the emotional curve to obtain an emotional tendency distribution probability; and an emotional tendency corresponding to a value greater than or equal to a preset threshold in the emotional tendency distribution probability is taken as the emotional feature of the user.

5. An electronic device, comprising: The first obtaining module is further configured to obtain a voiceprint feature included in the call recording; and determine the voice data of the user in the call recording according to the voiceprint feature and a preset voiceprint feature of the customer service. The computer program is executed by the processor to implement the method in any one of claims 1-3. The computer program is executed by the processor to implement the method in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer program is executed by the processor to implement the method in any one of claims 1-3.

7. A computer program product comprising a computer program, characterized in that, ​

Images