An emotion recognition method and device in a banking scenario

Abstract

The application provides a bank business scene emotion recognition method and device, relates to the artificial intelligence technical field, and can be applied to the metaverse and the financial technology field.The method comprises the following steps: if it is judged that the bank business handled by a customer belongs to specific business, obtaining a plurality of biological feature data, a business channel, a business type and customer historical complaint information;obtaining each biological feature of the customer based on the plurality of biological feature data;obtaining a weight prediction feature;obtaining an emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion recognition model;obtaining a weight corresponding to each biological feature based on various biological features, the weight prediction feature and a weight prediction model;and obtaining a customer emotion recognition result according to the emotion recognition result corresponding to each biological feature and the weight corresponding to each biological feature.The bank business scene emotion recognition method and device provided in the application embodiment improve the accuracy of emotion recognition.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence technology, which can be used in the fields of metaverse and financial technology, and in particular to a method and device for emotion recognition in a bank business scenario. BACKGROUND

[0002] Emotion recognition can automatically identify the emotional state of an individual by obtaining physiological or non-physiological signals, including facial expressions, speech, heart rate, behavior, text, and physiological signal recognition. The emotional state of the user is determined by the above data.

[0003] Existing emotion recognition schemes usually rely on a certain biological feature, such as analyzing facial information, voice information, and other biological features to obtain the current emotional state of the user. However, in the bank business scenario, there are cases where the current emotional state of the user cannot be accurately reflected. Therefore, how to propose an emotion recognition method that can more accurately identify the emotional state of the customer when the customer is conducting bank business is an important issue to be solved in the field. SUMMARY

[0004] To solve the problems in the prior art, the embodiments of the present application provide a method and device for emotion recognition in a bank business scenario, which can at least partially solve the problems in the prior art.

[0005] In a first aspect, the present application provides a method for emotion recognition in a bank business scenario, comprising:

[0006] If it is determined that the bank business handled by the customer belongs to a specific business, then a plurality of biological feature data required by the customer to handle the bank business, a business channel, a business type, and customer historical complaint information are obtained;

[0007] Based on the plurality of biological feature data required by the customer to handle the bank business, the customer's each biological feature is obtained; and the business channel, the business type, and the customer historical complaint information are preprocessed to obtain a weight prediction feature;

[0008] Based on each biological feature of the customer and a corresponding emotion recognition model, an emotion recognition result corresponding to each biological feature of the customer is obtained; wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training;

[0009] Based on the various biological features of the customer, the weight prediction feature, and a weight prediction model, a weight corresponding to each biological feature is obtained; wherein the weight prediction model is obtained based on weight sample data;

[0010] According to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight, a customer emotion recognition result is obtained.

[0011] In a second aspect, the present application provides an emotion recognition device in a banking scenario, comprising:

[0012] An acquisition module is configured to acquire a plurality of biological feature data, a service channel, a service type and historical complaint information of a customer after determining that the banking service handled by the customer belongs to a specific service;

[0013] An obtaining module is configured to obtain each biological feature of the customer based on the plurality of biological feature data required by the customer to handle the banking service, and to preprocess the service channel, the service type and the historical complaint information of the customer to obtain a weight prediction feature;

[0014] A first recognition module is configured to obtain an emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion recognition model, wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training;

[0015] A weight prediction module is configured to obtain a weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and a weight prediction model, wherein the weight prediction model is obtained by training based on weight sample data;

[0016] A second recognition module is configured to obtain an emotion recognition result of the customer based on the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight of each biological feature.

[0017] In a third aspect, the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the emotion recognition method in the banking scenario as described in any of the above embodiments.

[0018] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the emotion recognition method in the banking scenario as described in any of the above embodiments.

[0019] In a fifth aspect, the present application provides a computer program product, comprising a computer program, wherein the computer program is executable by a processor to implement the emotion recognition method in the banking scenario as described in any of the above embodiments.

[0020] The emotion recognition method and device in the bank service scenario provided by the embodiments of the present application, after judging that the bank service handled by the customer belongs to a specific service, obtain various biological feature data required by the customer to handle the bank service, a service channel, a service type and customer historical complaint information; obtain each biological feature of the customer based on the various biological feature data required by the customer to handle the bank service; and preprocess the service channel, the service type and the customer historical complaint information to obtain a weight prediction feature; obtain an emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion recognition model; obtain a weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and a weight prediction model; and obtain a customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight, since the emotion recognition is performed by comprehensively considering various biological features, service information and historical complaint information, the accuracy of the emotion recognition is improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor. In the drawings:

[0022] Figure 1 is a flowchart of the emotion recognition method in the bank service scenario provided by the first embodiment of the present application.

[0023] Figure 2 is a flowchart of the emotion recognition method in the bank service scenario provided by the second embodiment of the present application.

[0024] Figure 3 is a flowchart of the emotion recognition method in the bank service scenario provided by the third embodiment of the present application.

[0025] Figure 4 is a flowchart of the emotion recognition method in the bank service scenario provided by the fourth embodiment of the present application.

[0026] Figure 5 is a flowchart of the emotion recognition method in the bank service scenario provided by the fifth embodiment of the present application.

[0027] Figure 6 is a flowchart of the emotion recognition method in the bank service scenario provided by the sixth embodiment of the present application.

[0028] Figure 7is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by a seventh embodiment of the application.

[0029] Figure 8 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by an eighth embodiment of the application.

[0030] Figure 9 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by a ninth embodiment of the application.

[0031] Figure 10 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by a tenth embodiment of the application.

[0032] Figure 11 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by an eleventh embodiment of the application.

[0033] Figure 12 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by a twelfth embodiment of the application.

[0034] Figure 13 is a structural schematic diagram of an electronic device provided by a thirteenth embodiment of the application. DETAILED DESCRIPTION

[0035] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, further detailed description of the embodiments of the present application will be given below with reference to the drawings. Herein, the illustrative embodiments of the present application and their descriptions are used to explain the present application, but are not regarded as limiting the present application. It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other at will. The acquisition, storage, use, processing, etc. of data in the technical solutions in the present application all comply with the relevant provisions of laws and regulations. The user information in the embodiments of the present application is obtained through legal and compliant ways, and the acquisition, storage, use, processing, etc. of the user information are authorized and agreed by the client.

[0036] In order to facilitate understanding of the technical solutions provided by the present application, the related contents of the technical solutions of the present application will be described first.

[0037] The traditional emotion recognition method often depends on a certain biological feature, ignores the correlation between different biological features, and does not consider the actual business scene when recognizing emotions. Therefore, the embodiment of the present application proposes an emotion recognition method in a bank business scene, collects various biological feature data such as facial expressions, tone volume, heart rate, blood pressure and pupil conditions required by customers when conducting bank business, and combines multi-dimensional information such as the type of business currently done by the customer, to comprehensively analyze the biological features of the customer in the bank business scene, and obtain the emotion recognition result of the customer, so as to improve the accuracy of customer emotion recognition.

[0038] The specific implementation process of the emotion recognition method in the bank business scene provided by the embodiment of the present application will be described below taking the server as an execution subject.

[0039] Figure 1 is a flowchart of the emotion recognition method in the bank business scene provided by the first embodiment of the present application, as Figure 1 indicated, the emotion recognition method in the bank business scene provided by the embodiment of the present application comprises:

[0040] S101, if it is judged that the bank business conducted by the customer belongs to a specific business, obtaining various biological feature data required by the customer when conducting bank business, business channel, business type and customer historical complaint information;

[0041] Specifically, the customer can conduct bank business through counter, online bank, mobile bank, self-service terminal and the like. The server will judge whether the bank business conducted by the customer belongs to a specific business, and if it belongs to a specific business, the current emotion of the customer needs to be recognized. The server will obtain various biological feature data required by the customer when conducting bank business, business channel and business type, and will also obtain customer historical complaint information. The biological feature data includes but is not limited to fingerprint, face, voice and the like, which are collected according to actual business needs. The business channel includes but is not limited to counter, self-service terminal, online bank, mobile bank, telephone bank and the like. The customer historical complaint information refers to the complaint information of the customer to the bank within a set time period.

[0042] For example, some business types of bank business can be specified as specific business, and if the bank business conducted by the customer belongs to the specified business type, the bank business conducted by the customer belongs to specific business.

[0043] S102, based on the various biological feature data required by the customer when conducting bank business, obtaining each biological feature of the customer; and preprocessing the business channel, business type and customer historical complaint information to obtain weight prediction features;

[0044] Specifically, for each biometric data required by the customer to conduct bank business, the server will process each biometric data to obtain each biometric data of the customer. The server will preprocess the business channel, business type and customer historical complaint information to obtain the weight prediction feature. The preprocessing includes but is not limited to one-hot encoding and feature extraction.

[0045] For example, for the customer's fingerprint data, fingerprint features can be obtained; for the customer's face picture, face features can be obtained; for the customer's voice data, voiceprint features can be obtained.

[0046] For example, for the business channel and the business type, one-hot encoding can be used. For example, the business channel: counter, self-service terminal, online bank, mobile bank and telephone bank, which are respectively coded as 00001, 00010, 00100, 01000 and 10000.

[0047] For example, the customer historical complaint information is usually a text description, which can be converted into numerical information with the help of natural language processing technology, such as keyword extraction or more complex sentiment analysis technology. For example, the complaint frequency can be recorded, or it can be analyzed that the main problem of the complaint is service attitude, business error, etc., and these information is coded.

[0048] S103, based on each biometric feature of the customer and the corresponding emotion recognition model, obtaining the emotion recognition result corresponding to each biometric feature of the customer; wherein the emotion recognition model corresponding to each biometric feature is obtained by pre-training;

[0049] Specifically, the server inputs each biometric feature of the customer into the emotion recognition model corresponding to each biometric feature, and can obtain the emotion recognition result corresponding to each biometric feature of the customer. The emotion recognition result corresponding to each biometric feature can include the probability of the customer being in each emotional state. The emotion recognition model corresponding to each biometric feature is obtained by pre-training.

[0050] For example, based on the customer's face picture, facial expression features can be obtained, and the emotion recognition model corresponding to the facial expression features can be trained through the facial expression features.

[0051] Based on the customer's voice data, voiceprint features can be obtained, and the emotion recognition model corresponding to the voiceprint features can be trained through the voiceprint features.

[0052] The physiological indicators such as heart rate, respiratory rate, blood pressure, skin conductance, muscle tension and body temperature can also be used to train the corresponding emotion recognition model respectively.

[0053] S104, obtain the weight corresponding to each biological feature of the customer based on the various biological features of the customer, the weight prediction feature, and the weight prediction model, wherein the weight prediction model is obtained based on weight sample data;

[0054] Specifically, the server inputs the various biological features of the customer and the weight prediction feature into the weight prediction model, and can output the weight corresponding to each biological feature of the customer. The weight prediction model is obtained based on weight sample data.

[0055] S105, obtain the customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the weight corresponding to each biological feature.

[0056] Specifically, the server can obtain the customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the weight corresponding to each biological feature.

[0057] For example, the emotion recognition results A, B and C corresponding to three biological features of the customer are obtained, and the weights corresponding to each are a, b and c respectively. The emotion recognition result corresponding to each biological feature includes the probability of the customer being in five different emotional states. For an emotional state, the probability in the emotion recognition result A is x, the probability in the emotion recognition result B is y, and the probability in the emotion recognition result C is z, so the total probability P1 of the emotional state is ax+by+cz. Similarly, the total probabilities of the other four emotional states P2, P3, P4 and P5 can be obtained. By comparing P1, P2, P3, P4 and P5, the emotional state corresponding to the maximum value is taken as the customer emotion recognition result.

[0058] The emotion recognition method in the bank business scenario provided by the embodiment of the application obtains the various biological feature data required by the customer to handle the bank business, the business channel, the business type and the customer historical complaint information after judging that the bank business handled by the customer belongs to a specific business; obtains each biological feature of the customer based on the various biological feature data required by the customer to handle the bank business; and pre-processes the business channel, the business type and the customer historical complaint information to obtain a weight prediction feature; obtains the emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion recognition model; obtains the weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and the weight prediction model; and obtains the customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the weight corresponding to each biological feature. Since the emotion recognition is performed by comprehensively considering the various biological features, the business information and the historical complaint information, the accuracy of the emotion recognition is improved.

[0059] Figure 2is a flowchart of an emotion recognition method in a bank service scenario provided by a second embodiment of the present application, as shown in Figure 2 Further, based on the above embodiments, the weight prediction model is obtained by training based on the weight sample data, and the weight prediction model includes:

[0060] S201, obtain weight sample data, the weight sample data includes a plurality of data samples, each data sample includes various biological feature training data and weight feature training data of a customer handling bank business;

[0061] Specifically, the historical data corresponding to the bank business can be collected, and the weight sample data is obtained from the historical data corresponding to the bank business. The weight sample data includes a plurality of data samples. Each data sample includes various biological feature training data and weight feature training data of a customer handling bank business. The emotion label can include satisfaction, normal and dissatisfaction, which is set according to actual needs, and the embodiment of the present application is not limited.

[0062] S202, according to each data sample and the original model, a weight prediction model is trained and obtained.

[0063] Specifically, the server trains the original model based on the weight sample data, and can train and obtain the weight prediction model. The original model can adopt a neural network model, which is selected according to actual needs, and the embodiment of the present application is not limited. The specific training process of the model is prior art, which will not be described here.

[0064] Figure 3 is a flowchart of an emotion recognition method in a bank service scenario provided by a third embodiment of the present application, as shown in Figure 3 Further, based on the above embodiments, the weight sample data includes:

[0065] S301, based on the historical data corresponding to the bank business, various biological feature data, business channel and business type of a customer handling bank business are obtained, and historical complaint information corresponding to the bank business is obtained;

[0066] Specifically, the historical data corresponding to the bank business, i.e. the historical data of each customer handling various bank businesses, including various biological feature data, business channel and business type provided by the customer handling bank business, can be collected. From the historical data corresponding to the bank business, various biological feature data, business channel and business type of a customer handling bank business can be obtained. The complaint information of each customer on each bank business in the historical time period can be collected as the historical complaint information corresponding to the bank business. The historical time period is set according to actual needs, and the embodiment of the present application is not limited.

[0067] S302, obtain each kind of biological feature training data of each customer based on each kind of biological feature data of the customer handling the bank business; and preprocess the business channel, the business type and the historical complaint information of the customer of each customer handling the bank business to obtain the weight feature training data of each customer.

[0068] Specifically, the server can obtain each kind of biological feature training data of each customer by performing feature processing on each kind of biological feature data of the customer handling the bank business. The server can obtain the weight feature training data of each customer by preprocessing the business channel, the business type and the historical complaint information of the customer of each customer handling the bank business.

[0069] S303, taking each kind of biological feature training data and the weight feature training data of each customer as one data sample of the weight sample data.

[0070] Specifically, the server takes each kind of biological feature training data and the weight feature training data of each customer as one data sample, and each data sample constitutes the weight sample data.

[0071] Figure 4 is a flowchart of the emotion recognition method in the bank business scenario provided by the fourth embodiment of the present application, as shown in Figure 4 Further, the bank business handled by the customer belongs to a specific business if it is determined that the bank business handled by the customer belongs to a specific business, as shown in the above embodiments.

[0072] S401, obtaining the biological feature types and the bank business data corresponding to the bank business handled by the customer;

[0073] Specifically, the server can obtain the biological feature types corresponding to the bank business handled by the customer and the bank business data corresponding to the bank business handled by the customer when the customer handles the bank business. The bank business data includes but is not limited to the business type, the customer's business completion condition, the business amount, the transaction frequency, etc.

[0074] For example, for bank business, the business codes of each category of bank business can be traversed to find the code blocks related to biological feature recognition, such as biological feature data processing and calling. In the code blocks, the biological feature recognition related code blocks are usually marked with specific functions or keywords. By counting the specific functions or keywords related to biological feature recognition, the biological feature types corresponding to the bank business, i.e. the number of biological features required to handle the bank business, can be counted, and the counted biological feature types corresponding to the bank business are stored.

[0075] S402, if it is known that the biological feature type corresponding to the bank business handled by the customer is greater than 1, obtaining a risk prediction result of the bank business handled by the customer according to the bank business data and a risk prediction model, wherein the risk prediction model is obtained based on bank historical business data and corresponding risk labels.

[0076] Specifically, the server determines whether the biological feature type corresponding to the bank business handled by the customer is greater than 1. If the biological feature type corresponding to the bank business handled by the customer is greater than 1, the risk of the bank business is predicted according to the bank business data and the risk prediction model, and a risk prediction result of the bank business handled by the customer is obtained. The risk prediction result is a risk business or a non-risk business. The risk prediction model is obtained based on bank historical business data and corresponding risk labels.

[0077] S403, if the risk prediction result of the bank business handled by the customer is a risk business, determining that the bank business handled by the customer belongs to a specific business.

[0078] Specifically, if the risk prediction result of the bank business handled by the customer is a risk business, the bank business handled by the customer belongs to a specific business, and emotion recognition needs to be performed on the customer. For a non-risk business, the customer can not be subjected to emotion recognition, so as to speed up the processing efficiency of the bank business.

[0079] Through the risk prediction model, it can be determined which business needs to be paid attention to by the bank without manually formulating rules, and whether a new business belongs to a risk business can also be identified as the bank system is updated.

[0080] Figure 5 is a flowchart of an emotion recognition method in a bank business scenario provided by the fifth embodiment of the present application, as shown in Figure 5 Based on the above embodiments, further, the risk prediction model is obtained based on bank historical business data and corresponding risk labels, including:

[0081] S501, obtaining bank historical business data and corresponding risk labels;

[0082] Specifically, bank historical business data can be collected, and the bank historical business data includes business type, customer business completion condition, business amount, transaction frequency and the like. According to actual needs, selection is performed, and the embodiments of the present application are not limited. The bank historical business data includes a plurality of business sample data and a risk label corresponding to each business sample data, and the risk label can be obtained by manual annotation.

[0083] For example, a bank transaction amount is greater than a set amount and is a payment transaction, which can be marked as a risk transaction.

[0084] S502, based on the bank historical transaction data and the initial model, training obtains a risk prediction model.

[0085] Specifically, the server trains the model based on the bank historical transaction data and the initial model, and can train the risk prediction model. Wherein, the initial model can adopt a deep neural network model. The specific training process of the model is prior art, which will not be described here.

[0086] Wherein, for the bank historical transaction data, a self-encoder or a clustering algorithm can be used to extract key features. For example, for all payment transactions with an amount exceeding a certain amount, there may be similar risk characteristics.

[0087] Figure 6 is the flowchart of the emotion recognition method in the bank business scenario provided by the sixth embodiment of the application, as shown in Figure 6 Based on the above embodiments, further, before acquiring the multiple biological feature data required by the customer to handle the bank business, it further includes:

[0088] S601, receiving the hardware list reported by the client;

[0089] Specifically, before collecting the multiple biological feature data of the customer, the client can report the hardware list to the server, and the server can receive the hardware list. The hardware list includes but is not limited to camera, microphone, fingerprint sensor, etc. The client includes but is not limited to mobile terminal, desktop, self-service terminal, etc.

[0090] S602, if the hardware devices in the hardware list do not match the multiple biological features required by the customer to handle the bank business, send a prompt information of lacking biological feature acquisition hardware to the client.

[0091] Specifically, the server judges whether the hardware devices in the hardware list match the multiple biological features required by the customer to handle the bank business, that is, whether the multiple biological feature data required by the customer to handle the bank business can be collected through the hardware devices in the hardware list. If the hardware devices in the hardware list do not match the multiple biological features required by the customer to handle the bank business, it means that all biological feature data required by the customer to handle the bank business cannot be obtained, and the server will send a prompt information of lacking biological feature acquisition hardware to the client to prompt the customer that the current client cannot handle the business.

[0092] By judging whether the hardware of the client matches the biological characteristics required by the bank business or not, the bank business that cannot be handled due to hardware reasons is prompted in time, and the business handling experience of the client is improved.

[0093] Further, on the basis of the above-mentioned embodiments, the emotion recognition method in the bank business scenario provided by the embodiments of the present application further comprises:

[0094] Synchronizing the customer emotion recognition result to a background system.

[0095] Specifically, after obtaining the customer emotion recognition result, the server can synchronize the customer emotion recognition result to the background system, so that the background system provides better service for the customer by using the customer emotion recognition result.

[0096] For example, the banking industry has launched various functions with a meta-universe label on the mobile client, creates exclusive roles for customers, and handles business in a virtual space. However, since the customers and customer service personnel in the meta-universe are virtual roles, and the virtual roles are not associated with customer emotions, it is more difficult for the bank customer service personnel to recognize the negative emotions such as anxiety and dissatisfaction of the customers due to the inability of face-to-face communication between people, so that the customers complain due to the inability to provide assistance in time. Synchronizing the customer emotion recognition result to the background system of the meta-universe service and writing the customer emotion recognition result into the customer meta-universe role can help the customer service personnel in the meta-universe to recognize the customer emotions through the meta-universe role and improve the service.

[0097] Figure 7 is a structural schematic diagram of an emotion recognition device in a bank business scenario provided by the seventh embodiment of the present application, as Figure 7 shown, the emotion recognition device in the bank business scenario provided by the embodiments of the present application comprises an acquisition module 701, an obtaining module 702, a first recognition module 703, a weight prediction module 704, and a second recognition module 705, wherein:

[0098] The acquisition module 701 is configured to acquire a plurality of biological feature data, a service channel, a service type and historical complaint information of a customer after determining that the bank service handled by the customer belongs to a specific service; the obtaining module 702 is configured to obtain each biological feature of the customer based on the plurality of biological feature data required by the customer to handle the bank service, and preprocess the service channel, the service type and the historical complaint information of the customer to obtain a weight prediction feature; the first identification module 703 is configured to obtain an emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion recognition model; wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training; the weight prediction module 704 is configured to obtain a weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and a weight prediction model; wherein the weight prediction model is obtained by training based on weight sample data; and the second identification module 705 is configured to obtain a customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight of each biological feature.

[0099] Specifically, the customer can handle the bank service through a counter, online banking, mobile banking, a self-service terminal and the like. The acquisition module 701 determines whether the bank service handled by the customer belongs to a specific service, and if it belongs to a specific service, the current emotion of the customer needs to be identified. The acquisition module 701 acquires a plurality of biological feature data, a service channel and a service type required by the customer to handle the bank service, and also acquires historical complaint information of the customer. The biological feature data includes but is not limited to fingerprint, face, voice and the like, which are collected according to actual business needs. The service channel includes but is not limited to counter, self-service terminal, online banking, mobile banking, telephone banking and the like. The historical complaint information of the customer refers to the complaint information of the customer to the bank within a set time period.

[0100] For each biological feature data required by the customer to handle the bank service, the obtaining module 702 processes each biological feature data to obtain each biological feature data of the customer. The obtaining module 702 preprocesses the service channel, the service type and the historical complaint information of the customer to obtain a weight prediction feature. The preprocessing includes but is not limited to one-hot encoding and feature extraction.

[0101] The first identification module 703 inputs each biological feature of the customer into the emotion recognition model corresponding to each biological feature to obtain an emotion recognition result corresponding to each biological feature of the customer. The emotion recognition result corresponding to each biological feature can include the probability of the customer being in each emotional state. The emotion recognition model corresponding to each biological feature is obtained by pre-training.

[0102] The weight prediction module 704 inputs various biological features of the customer and the weight prediction feature into a weight prediction model, and can output a weight corresponding to each biological feature of the customer. The weight prediction model is obtained by training based on weight sample data.

[0103] The second identification module 705 obtains the customer emotion identification result according to the emotion identification result corresponding to each biological feature of the customer and the respective corresponding weight.

[0104] The emotion identification method and device in the bank business scenario provided by the embodiment of the present application, after judging that the bank business handled by the customer belongs to a specific business, obtaining a plurality of biological feature data required by the customer to handle the bank business, a business channel, a business type and customer historical complaint information; obtaining each biological feature of the customer based on the plurality of biological feature data required by the customer to handle the bank business; and preprocessing the business channel, the business type and the customer historical complaint information to obtain a weight prediction feature; obtaining an emotion identification result corresponding to each biological feature of the customer based on each biological feature of the customer and a corresponding emotion identification model; obtaining a weight corresponding to each biological feature based on each biological feature of the customer, the weight prediction feature and a weight prediction model; and obtaining a customer emotion identification result according to the emotion identification result corresponding to each biological feature of the customer and the respective corresponding weight. Since the emotion identification is performed by comprehensively considering a plurality of biological features, business information and historical complaint information, the accuracy of the emotion identification is improved.

[0105] Figure 8 is a structural schematic diagram of the emotion identification device in the bank business scenario provided by the eighth embodiment of the present application, as Figure 8 shown, on the basis of the above-mentioned embodiments, further, the emotion identification device in the bank business scenario provided by the embodiment of the present application further includes a first sample obtaining module 706 and a first training module 707, wherein:

[0106] The first sample obtaining module 706 is used to obtain weight sample data, and the weight sample data includes a plurality of data samples, each data sample includes various biological feature training data and weight feature training data of the customer handling the bank business; the first training module 707 is used to train the weight prediction model according to each data sample and an original model.

[0107] Figure 9 is a structural schematic diagram of the emotion identification device in the bank business scenario provided by the ninth embodiment of the present application, as Figure 9 shown, on the basis of the above-mentioned embodiments, further, the first sample obtaining module 706 includes a first obtaining unit 7061, a second obtaining unit 7062 and an as unit 7063, wherein:

[0108] The first obtaining unit 7061 is configured to obtain various biological feature data of the customer, a service channel and a service type of the bank service, and obtain historical complaint information corresponding to the bank service based on historical data corresponding to the bank service; the second obtaining unit 7062 is configured to obtain biological feature training data of each customer based on each biological feature data of the customer; and the service channel, the service type and the historical complaint information corresponding to the bank service of each customer are preprocessed to obtain weight feature training data of each customer; and the unit 7063 is configured to take the various biological feature training data and the weight feature training data of each customer as a data sample of the weight sample data.

[0109] Figure 10 is a structural schematic diagram of an emotion recognition device in a bank service scenario provided by the tenth embodiment of the present application, as shown in the above embodiments, further, the obtaining module 701 comprises an obtaining unit 7011, a judging unit 7012 and a determining unit 7013, wherein: Figure 10

[0110] The obtaining unit 7011 is configured to obtain biological feature types corresponding to the bank service and bank service data; the judging unit 7012 is configured to obtain a risk prediction result of the bank service of the customer according to the bank service data and a risk prediction model if it is judged that the biological feature types corresponding to the bank service of the customer are greater than 1; wherein the risk prediction model is obtained by training based on historical bank service data and corresponding risk labels; and the determining unit 7013 is configured to determine that the bank service of the customer belongs to a specific service if the risk prediction result of the bank service of the customer is a risk service.

[0111] Figure 11 is a structural schematic diagram of an emotion recognition device in a bank service scenario provided by the eleventh embodiment of the present application, as shown in the above embodiments, further, the bank service scenario emotion recognition device provided by the embodiment of the present application further comprises a second sample obtaining module 708 and a second training module 709, wherein: Figure 11

[0112] The second sample obtaining module 708 is configured to obtain historical bank service data and corresponding risk labels; and the second training module 709 is configured to train a risk prediction model based on the historical bank service data and an initial model.

[0113] Figure 12 is a structural schematic diagram of an emotion recognition device in a bank service scenario provided by the twelfth embodiment of the present application, as shown in the above embodiments, further, the bank service scenario emotion recognition device provided by the embodiment of the present application further comprises a second sample obtaining module 708 and a second training module 709, wherein: Figure 12 ​​As shown, on the basis of each of the above embodiments, further, the emotion recognition device under the bank business scenario provided by the embodiment of the application further includes a receiving module 710 and a prompt module 711, wherein:

[0114] The receiving module 710 is used for receiving a hardware list reported by a client; and the prompt module 711 is used for sending prompt information of lacking biological characteristic acquisition hardware to the client if a hardware device in the hardware list does not match a plurality of biological characteristics required by a customer to handle bank business.

[0115] On the basis of each of the above embodiments, further, the emotion recognition device under the bank business scenario provided by the embodiment of the application further includes a synchronization module, wherein:

[0116] The synchronization module is used for synchronizing the customer emotion recognition result to a background system.

[0117] The embodiment of the server provided by the embodiment of the application can be specifically used for executing the processing flow of each of the above method embodiments, and the function thereof will not be repeated here, and the detailed description can be referred to the detailed description of the above method embodiments.

[0118] It should be noted that the emotion recognition method and device under the bank business scenario provided by the embodiment of the application can be used in the financial field, and can also be used in any technical field other than the financial field, and the application field of the emotion recognition method and device under the bank business scenario is not limited by the embodiment of the application.

[0119] Figure 13 is the entity structure schematic diagram of the electronic equipment provided by the thirteenth embodiment of the application, like Figure 13As shown, the electronic device can include a processor 1301, a communications interface 1302, a memory 1303, and a communications bus 1304, wherein the processor 1301, the communications interface 1302, and the memory 1303 complete mutual communication through the communications bus 1304. The processor 1301 can invoke a logical instruction in the memory 1303 to execute the following method: if it is judged that the bank business handled by the customer belongs to a specific business, obtaining a plurality of biological characteristic data required by the customer to handle the bank business, a business channel, a business type, and customer historical complaint information; obtaining each biological characteristic of the customer based on the plurality of biological characteristic data required by the customer to handle the bank business; and preprocessing the business channel, the business type, and the customer historical complaint information to obtain a weight prediction feature; obtaining an emotional recognition result corresponding to each biological characteristic of the customer based on each biological characteristic of the customer and a corresponding emotional recognition model; wherein the emotional recognition model corresponding to each biological characteristic is obtained by pre-training; obtaining a weight corresponding to each biological characteristic based on the various biological characteristics of the customer, the weight prediction feature, and a weight prediction model; wherein the weight prediction model is obtained by training based on weight sample data; and obtaining a customer emotional recognition result according to the emotional recognition result corresponding to each biological characteristic of the customer and the corresponding weight.

[0120] In addition, the logical instructions in the memory 1303 described above can be implemented in the form of a software function unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0121] The embodiment discloses a computer program product, the computer program product comprises a computer program stored on a computer readable storage medium, the computer program comprises program instructions, when the program instructions are executed by a computer, the computer can execute the method provided by each method embodiment, for example, comprising: if it is judged that the bank business handled by the customer belongs to a specific business, obtaining a plurality of biological feature data required by the customer to handle the bank business, a business channel, a business type and customer historical complaint information; obtaining each biological feature of the customer based on the plurality of biological feature data required by the customer to handle the bank business; and preprocessing the business channel, the business type and the customer historical complaint information to obtain a weight prediction feature; obtaining the emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and the corresponding emotion recognition model; wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training; obtaining the weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and the weight prediction model; wherein the weight prediction model is obtained based on weight sample data; and obtaining the customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight.

[0122] The embodiment provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program enables the computer to execute the method provided by each method embodiment, for example, comprising: if it is judged that the bank business handled by the customer belongs to a specific business, obtaining a plurality of biological feature data required by the customer to handle the bank business, a business channel, a business type and customer historical complaint information; obtaining each biological feature of the customer based on the plurality of biological feature data required by the customer to handle the bank business; and preprocessing the business channel, the business type and the customer historical complaint information to obtain a weight prediction feature; obtaining the emotion recognition result corresponding to each biological feature of the customer based on each biological feature of the customer and the corresponding emotion recognition model; wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training; obtaining the weight corresponding to each biological feature based on the various biological features of the customer, the weight prediction feature and the weight prediction model; wherein the weight prediction model is obtained based on weight sample data; and obtaining the customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight.

[0123] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0124] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for performing the function specified by the flowchart illustrations and / or block diagrams block or blocks.

[0125] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for performing the function specified by the flowchart illustrations and / or block diagrams block or blocks.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart illustrations and / or block diagrams block or blocks. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams block or blocks. Figure 1 means for performing the function specified by the flowchart illustrations and / or block diagrams block or blocks.

[0127] In the description of the specification, the description of the terms "one embodiment", "one specific embodiment", "some embodiments", "for example", "an example", "a specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or more embodiments or examples.

[0128] The specific embodiments described above further illustrate the purposes, technical solutions and beneficial effects of the present application. It should be understood that the above description is only for specific embodiments of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An emotion recognition method for banking business scenarios, characterized in that, The method comprises the following steps: If it is determined that the bank service handled by the customer belongs to a specific service, obtaining a plurality of biological feature data required by the customer to handle the bank service, a service channel, a service type and customer historical complaint information; Based on the plurality of biological feature data required by the customer to handle the bank service, obtaining each biological feature of the customer; and preprocessing the service channel, the service type and the customer historical complaint information to obtain a weight prediction feature; Based on each biological feature of the customer and a corresponding emotion recognition model, obtaining an emotion recognition result corresponding to each biological feature of the customer; wherein the emotion recognition model corresponding to each biological feature is obtained by pre-training; Based on the various biological features of the customer, the weight prediction feature and a weight prediction model, obtaining a weight corresponding to each biological feature; wherein the weight prediction model is obtained by training based on weight sample data; Obtaining a customer emotion recognition result according to the emotion recognition result corresponding to each biological feature of the customer and the corresponding weight; Training the weight prediction model based on the weight sample data comprises: Obtaining weight sample data, wherein the weight sample data comprises a plurality of data samples, and each data sample comprises various biological feature training data and weight feature training data of the customer handling the bank service; Training the weight prediction model based on each data sample and an original model; The obtaining of the weight sample data comprises: Based on the historical data corresponding to the bank service, obtaining the various biological feature data, the service channel and the service type of the customer handling the bank service, and obtaining the historical complaint information corresponding to the bank service; Based on each biological feature data of the customer handling the bank service, obtaining each biological feature training data of each customer; preprocessing the service channel, the service type and the historical complaint information corresponding to the bank service of each customer handling the bank service to obtain weight feature training data of each customer; Taking the various biological feature training data and the weight feature training data of each customer as a data sample of the weight sample data.

2. The method of claim 1, wherein, The determination that the bank service handled by the customer belongs to a specific service comprises: Obtaining the biological feature type corresponding to the bank service handled by the customer and bank service data; If it is determined that the biological feature type corresponding to the bank service handled by the customer is greater than 1, obtaining a risk prediction result of the bank service handled by the customer according to the bank service data and a risk prediction model; wherein the risk prediction model is obtained by training based on bank historical service data and a corresponding risk label; If the risk prediction result of the bank service handled by the customer is a risk service, it is determined that the bank service handled by the customer belongs to a specific service.

3. The method of claim 2, wherein, Training the risk prediction model based on the bank historical service data and the corresponding risk label comprises: Obtaining the bank historical service data and the corresponding risk label; Training the risk prediction model based on the bank historical service data and an initial model.

4. The method of claim 1, wherein, Before obtaining the plurality of biological feature data required by the customer to handle the bank service, the method further comprises: Receiving a hardware list reported by a client; If the hardware device in the hardware list does not match the multiple biological characteristics required by the customer for bank business, a prompt message of lacking biological characteristic collection hardware is sent to the client.

5. The method according to any one of claims 1 to 4, characterized in that, Also include: Synchronize the customer emotion recognition result to the background system.

6. An emotion recognition apparatus in a banking scenario, configured to perform the method of any one of claims 1-5, characterized in that, The device includes: The acquisition module is configured to, after determining that the bank business handled by the customer belongs to a specific business, acquire multiple biological characteristic data required by the customer for bank business, a business channel, a business type, and customer historical complaint information; The obtaining module is configured to, based on the multiple biological characteristic data required by the customer for bank business, obtain each biological characteristic of the customer; and preprocess the business channel, the business type, and the customer historical complaint information to obtain weight prediction features; The first identification module is configured to, based on each biological characteristic of the customer and a corresponding emotion recognition model, obtain an emotion recognition result corresponding to each biological characteristic of the customer; wherein the emotion recognition model corresponding to each biological characteristic is obtained by pre-training; The weight prediction module is configured to, based on the various biological characteristics of the customer, the weight prediction features, and a weight prediction model, obtain a weight corresponding to each biological characteristic; wherein the weight prediction model is obtained by training based on weight sample data; The second identification module is configured to, based on the emotion recognition result corresponding to each biological characteristic of the customer and the corresponding weight, obtain a customer emotion recognition result.

7. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the method of any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the method of any one of claims 1 to 5.

9. A computer program product comprising computer programs or instructions, characterized in that, The computer program or instructions are executed by the processor to realize the steps of the method according to any one of claims 1 to 5.

Images