Recommended methods, systems, electronic devices, and media

Abstract

The present application provides a recommendation method that can be used in the field of artificial intelligence technology. The recommendation method includes: obtaining the service feature information of m account managers and the user portrait of the customer; generating a first recommendation list based on the customer's user portrait and the service feature information of the m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer; using the recommendation index model to calculate the recommendation index of each account manager in the first recommendation list; sorting the n account managers based on the recommendation index to generate a second recommendation list; and recommending the first-ranked account manager to the customer based on the second recommendation list. The recommendation method of the present application further sorts the account managers on the basis of the first recommendation list, and finally recommends the account manager ranked first on the second recommendation list. This method has a high degree of fit with customers and is close to the real demands of customers. It can not only effectively improve customer satisfaction, but also improve the marketing performance of account managers.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to a recommendation method, system, electronic device, medium, and program product. Background Art

[0002] Currently, the recommendation methods and techniques for account managers in the financial services industry are mainly based on customer portraits and account manager portraits, and use artificial intelligence algorithms for calculation and recommendation. However, they do not fully consider the customer's experience and feelings during the service process, and there is a problem of low recommendation accuracy. Summary of the Invention

[0003] This application aims to solve at least one of the technical problems existing in the prior art.

[0004] For example, this application provides a recommendation method that uses digital holographic technology to obtain the customer's three-dimensional service experience, and then uses artificial intelligence technology to analyze the customer's emotional changes during the experience, and then optimize the model's recommendation algorithm to meet the customer's actual needs and improve the customer's service experience.

[0005] In order to solve the above problems, the first aspect of the present application provides a recommendation method, comprising the following steps:

[0006] Get the service feature information of m account managers;

[0007] Obtaining a user portrait of at least one customer, wherein at least one of the customers has completed collection authorization;

[0008] Generate a first recommendation list based on the user profile of the at least one customer and the service feature information of the m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer, where m≥n, and both m and n are greater than or equal to 1;

[0009] Calculating a recommendation index for each account manager in the first recommendation list using a recommendation index model, wherein the recommendation index is used to represent the degree of compatibility with the client;

[0010] Sort the n account managers based on the recommendation index to generate a second recommendation list;

[0011] According to the second recommendation list, a first-ranked account manager is recommended to the client.

[0012] According to the recommendation method of this application, a first recommendation list suitable for recommendation is generated based on the user profile and the account manager profile. Based on this first recommendation list, the account managers on the first recommendation list are further ranked by recommendation index to generate a second recommendation list, and the account manager ranked first on the second recommendation list is finally recommended. This method is highly compatible with customers and closely reflects their real needs. It can not only effectively improve customer satisfaction, but also enhance the marketing performance of account managers.

[0013] Furthermore, before calculating the recommendation index of each account manager in the first recommendation list using the recommendation index model, the method further includes:

[0014] Obtain a historical service data of the customer within time period t;

[0015] According to the a historical service data, obtaining the emotional fluctuation information of the customer in each historical service data;

[0016] Obtaining post-service indicator information corresponding to the customer's emotional fluctuation information;

[0017] The recommendation index model is established according to the post-service indicator information and the emotion fluctuation information.

[0018] Furthermore, the historical service data includes holographic data and voice data.

[0019] Furthermore, according to the a historical service data, the emotional fluctuation information of the customer in each historical service data is obtained, including:

[0020] extracting the holographic data and the voice data for the i-th time, where i belongs to a;

[0021] extracting keywords from the voice data;

[0022] Based on the keywords, the i-th service is divided into multiple service stages;

[0023] Obtain holographic images of each service stage;

[0024] Obtaining at least one emotional feature value of the customer in the holographic image using an emotion recognition model;

[0025] Calculate the total emotion score of each service stage based on the emotion feature value;

[0026] The total emotion score of each service stage is analyzed to obtain the customer's emotion fluctuation information.

[0027] Furthermore, using an emotion recognition model, at least one emotion feature value of the customer in each service stage is obtained, including:

[0028] Build an emotion recognition model that includes multiple emotion features;

[0029] Inputting holographic images into the emotion recognition model;

[0030] Analyze the holographic image and output the emotional feature value of each emotional feature.

[0031] Furthermore, the total emotion score of each service stage is calculated based on the emotion feature value, including:

[0032] Classify each emotional characteristic as either positive or negative;

[0033] Positive emotions are assigned positive values, and negative emotions are assigned negative values;

[0034] The emotion feature values ​​of each service stage are added together.

[0035] Furthermore, the total emotion score of each service stage is analyzed to obtain the customer's emotion fluctuation information, including:

[0036] Calculate the slope based on the total emotion scores of two adjacent service stages;

[0037] The slopes are added to obtain the emotion fluctuation information.

[0038] Furthermore, the emotion fluctuation information includes positive fluctuation, negative fluctuation and smooth fluctuation.

[0039] Furthermore, the post-service indicator information includes post-service behavior information and an indicator score generated according to the post-service behavior information.

[0040] Furthermore, it also includes:

[0041] According to the service evaluations fed back by the customers, when the matching degree of the first-ranked account manager is lower than a preset range, the service feature information of the account manager is re-analyzed and the first recommendation list is updated.

[0042] Furthermore, the service feature information includes at least one of the account manager's years of experience, resume information, marketing performance and business scope.

[0043] Furthermore, the user portrait includes at least one of the customer's basic information, risk preferences and historical purchase information.

[0044] The second aspect of the present application provides a recommendation system, including: a first acquisition module, the first acquisition module is used to obtain service feature information of m account managers; a second acquisition module, the second acquisition module is used to obtain the user portrait of the customer; a first generation module, the first generation module is used to generate a first recommendation list based on the user portrait of the customer and the service feature information of the m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer, m≥n, and both m and n are greater than or equal to 1; a calculation module, the calculation module is used to calculate the recommendation index of each account manager in the first recommendation list using a recommendation index model; a second generation module, the second generation module is used to sort the n account managers based on the recommendation index to generate a second recommendation list; and a recommendation module, the recommendation module is used to recommend the first-ranked account manager to the customer based on the second recommendation list.

[0045] The third aspect of the present application provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned recommendation method.

[0046] The fourth aspect of the present application further provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above-mentioned recommendation method.

[0047] The fifth aspect of the present application also provides a computer program product, including a computer program, which implements the above-mentioned recommendation method when executed by a processor. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above contents and other objects, features and advantages of the present application will become more apparent through the following description of the embodiments of the present application with reference to the accompanying drawings, in which:

[0049] Figure 1 Schematically illustrates an application scenario diagram of a recommended method, apparatus, system, device, medium, and program product according to an embodiment of the present application;

[0050] Figure 2 A flowchart of a recommendation method according to an embodiment of the present application is schematically shown;

[0051] Figure 3 Schematically shows a flow chart of a method before using a recommendation index model according to an embodiment of the present application;

[0052] Figure 4 A flowchart of a method for obtaining emotion fluctuation information according to an embodiment of the present application is schematically shown;

[0053] Figure 5 A flowchart of a method for analyzing a holographic image using an emotion recognition model according to an embodiment of the present application is schematically shown;

[0054] Figure 6 A flowchart of a method for calculating the total emotion score for each service stage according to an embodiment of the present application is schematically shown;

[0055] Figure 7 Schematically shows a flow chart of a method for analyzing emotional fluctuation information according to an embodiment of the present application;

[0056] Figure 8 Schematically shows a structural diagram of a client manager recommendation device according to an embodiment of the present application;

[0057] Figure 9 A schematic diagram of a recommendation system according to an embodiment of the present application is shown; and

[0058] Figure 10 The block diagram schematically shows an electronic device suitable for implementing the recommendation method according to an embodiment of the present application. DETAILED DESCRIPTION

[0059] Hereinafter, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present application.

[0060] The terms used herein are only for describing specific embodiments and are not intended to limit this application. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0061] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0062] When expressions such as "at least one of A, B and C, etc." are used, they should generally be interpreted in accordance with the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0063] Currently, the recommendation methods and techniques for account managers in the financial services industry are mainly based on customer portraits and account manager portraits, and use artificial intelligence algorithms for calculation and recommendation. However, they do not fully consider the customer's experience and feelings during the service process, and there is a problem of low recommendation accuracy.

[0064] With the development of 5G communications, Internet technology and multimedia technology, some mobile terminals now have holographic projection communication capabilities, making it possible to use digital holography and 5G technology to collect three-dimensional customer information and optimize customer manager recommendation algorithms.

[0065] This application uses digital holographic technology to obtain the customer's three-dimensional service experience, mainly to obtain the customer's emotions and attitudes, and then uses artificial intelligence technology to optimize the algorithm of the customer manager recommendation model based on the obtained customer emotions and attitudes. The recommendation results calculated using this customer manager recommendation model are more in line with the customer's actual needs and can improve the customer's service experience.

[0066] It is important to note that digital holography in this application uses photoelectric sensors instead of dry plates to record holograms. The holograms are then stored in a computer, and the computer simulates the optical diffraction process to achieve holographic online processing of the recorded objects. Digital holography combines mathematical techniques with traditional Guangxu holography technology, using CCD (charge-coupled device) as the hologram recording medium and recreating the object light wavefront through computer numerical simulation of the optical diffraction process, thus creating realistic three-dimensional objects.

[0067] It is understandable that this application uses digital holographic technology to replay the service process between the account manager and the customer, and can use a three-dimensional method to show the customer's facial expressions and posture when being served, so as to optimize the account manager recommendation algorithm.

[0068] Figure 1 The following schematically illustrates an application scenario 100 of the recommended method, apparatus, system, device, and medium according to an embodiment of the present application. Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present application can be applied, to help those skilled in the art understand the technical content of the present application, but does not mean that the embodiments of the present application cannot be used in other devices, apparatuses, systems, environments or scenarios.

[0069] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a client terminal 101, a server 102, and client manager terminals 103, 104, and 105. The client terminal 101, the server 102, and the client manager terminals 103, 104, and 105 are connected via a network. The network may be provided via a wired or wireless connection medium.

[0070] The application scenario 100 may also include a customer 10 and customer managers 21, 22, and 23. Customer 10 is a customer who logs in to the customer terminal 101 using specific customer identity information (e.g., customer number). Customer managers 21, 22, and 23 are customer managers who log in to the customer manager terminals 103, 104, and 105 using specific customer manager identity information (e.g., customer manager marketing code), respectively.

[0071] Client 10 can use client terminal 101 to send a service request to server 102. After receiving the client's service request, server 102 can use the recommended method of the present application to determine that the account manager who provides service to client 10 is one of account managers 21, 22, and 23 (for example, account manager 21). In this way, server 102 can send information to account manager terminal 103 to establish a connection between account manager terminal 103 and client terminal 101, so that account manager 21 can provide the corresponding service to client 10.

[0072] It should be noted that the recommendation method provided in this application can generally be executed by the server 102. Accordingly, the recommendation system provided in this application can generally be set in the server 102. The recommendation method provided in this application can also be executed by a server or server cluster that is different from the server 102 and can communicate with the client terminal 101 and / or the customer manager terminals 103, 104, 105, and / or the server 102. Accordingly, the recommendation system provided in this application can also be set in a server or server cluster that is different from the server 102 and can communicate with the client terminal 101 and / or the customer manager terminals 103, 104, 105, and / or the server 102.

[0073] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0074] The following will be based on Figure 1 The scene described by Figures 2 to 7 The recommended method of the application embodiment is described in detail.

[0075] It should be noted that in the technical solution of this application, the acquisition, storage and application of customer personal information involved have all been collected after obtaining the consent of the person concerned, comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0076] Figure 2 The flowchart of the recommendation method according to an embodiment of the present application is schematically shown.

[0077] like Figure 2 As shown, this embodiment can be used for a method for customer manager recommendation, including operations S210 to S260.

[0078] In operation S210 , service feature information of m customer managers is obtained.

[0079] The database stores the service characteristic information of m account managers. The service characteristic information of each account manager is extracted from the database to establish the corresponding portrait information of each account manager. The portrait information is associated with the marketing code of the account manager so that the service characteristic information of the relevant account manager can be directly browsed after entering the marketing code.

[0080] The service characteristic information may include at least one of each account manager's years of experience, resume information, marketing achievements, and business scope. Of course, the service characteristic information is not limited to the above-mentioned types. Other service characteristic information that helps to portray the account manager's profile can also be used as content objects for acquisition.

[0081] In operation S220, a user portrait of at least one customer is obtained, wherein at least one customer has completed collection authorization.

[0082] After the customer agrees to collect and analyze their personal information, we extract the basic personal information they provided during registration and use it to create a user profile. Upon successful registration, the customer receives a unique customer number, which is associated with the user profile so that the customer's basic personal information can be directly retrieved by entering the customer number.

[0083] The user profile includes at least one of the following: basic customer information (e.g., age, gender, nationality, occupation, assets), risk preferences, and historical purchase information. Of course, the data and information that constitute the user profile is not limited to the aforementioned types; other data and information that contributes to the user profile can also be used as content to be acquired.

[0084] In operation S230, a first recommendation list is generated based on the user portrait of at least one customer and the service feature information of m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer, m≥n, and both m and n are greater than or equal to 1.

[0085] Based on the user profile and the account manager's service profile information, a first list of recommended services for the relevant customer is generated. This first list of recommended services can be obtained by building a model. A machine learning clustering algorithm is used to build a model based on the user profile and the account manager's service profile information. The customer and account manager's characteristics are used as training samples to obtain the first list of recommended services.

[0086] The main steps include:

[0087] 1) User portrait data analysis: User portrait data is used as customer feature data, and a machine learning clustering algorithm (K-means) is used to classify customers. Customers with similar features are grouped into the same cluster, and customers in the same cluster are considered part of the same customer group.

[0088] 2) Account manager portrait data analysis: Based on the characteristics of the account manager portrait, the account managers are classified using a clustering algorithm. Account managers in the same cluster have more similar characteristics and belong to the same service group.

[0089] 3) Analyze the mapping relationship between account managers and customer groups. This mapping relationship can be established by establishing a corresponding mapping for each customer the account manager has served, or establishing a corresponding mapping for each customer who has shared characteristics. Then, by calculating the number of mappings, a corresponding relationship between customer groups and service groups is established, resulting in a recommended list of account managers.

[0090] Through the above steps, we can select multiple suitable account managers based on a customer's user profile. This list of multiple account managers serves as the first recommendation list. This means that n out of m account managers will fit the customer's user profile and be suitable for serving this customer. Of course, some customers with similar user profiles can be grouped into the same cluster, corresponding to the same lists of multiple account managers. In other words, customers in the same cluster share the same first recommendation list.

[0091] In one embodiment, through data analysis of user profiles, customer groups are divided into customer group A, customer group B, customer group C,..., where individual customers in customer group A are represented as A1, A2, A3,...; through data analysis of account manager profiles, service groups are divided into service group X, service group Y, service group Z,..., where individual account managers in service group X are represented as X1, X2, X3,... . An aggregation algorithm is used to establish a mapping relationship between customer groups and service groups. The number of mapping relationships indicates the compatibility between the service group and the customer group. During the analysis of the number of mapping relationships, the service group with the highest degree of matching for each customer group can be determined based on the number of existing mapping relationships. For example, if service group X has a large number of mappings with customer group A, then service group X and customer group A are more closely matched.

[0092] It should be noted that the relationship between service groups and customer groups can be one-to-many, meaning that one account manager can be associated with multiple customers. For example, account manager X1 in service group X is suitable for recommending customers B1 and B2 in customer group B. Alternatively, one customer can be well-suited to being served by multiple account managers. For example, customer B1 in customer group B is highly compatible with account managers X1 and X2 in service group X. Alternatively, the relationship can be one-to-one, meaning that one account manager is only suitable for serving one customer.

[0093] Customers are divided into new and existing customers. New customers can be understood as those who have only registered but have not yet been served, while existing customers can be understood as those who have already been served. A customer group may consist entirely of existing customers, entirely of new users, or even partially of existing and new users. The principles for forming service groups and customer groups are similar. Due to the various methods for establishing mapping relationships, operation S230 is applicable to clustering, establishing associations, and generating the first recommendation list for all customer groups and all service groups.

[0094] In operation S240 , a recommendation index model is used to calculate the recommendation index of each account manager in the first recommendation list. The recommendation index is used to represent the degree of compatibility with the client.

[0095] The first recommendation list is generated based on the number of associations. For a customer, the first recommendation list can be understood as n account managers that match the customer.

[0096] In order to further screen the n account managers and select the account manager that is most suitable for the client, it is necessary to calculate the recommendation index of the n account managers in the first recommendation list based on the client, that is, to calculate the account manager with the highest fit with the client.

[0097] Further screening will help improve the fit between customers and account managers, and get closer to customers' real demands. It can not only effectively improve customer satisfaction, but also improve the marketing performance of account managers.

[0098] In operation S250 , the n account managers are ranked based on the recommendation index to generate a second recommendation list.

[0099] It can be understood that the second recommendation list is a recommendation list after n account managers have been sorted based on the first recommendation list.

[0100] In operation S260 , a first-ranked account manager is recommended to the client based on the second recommendation list.

[0101] After sorting the n account managers, a second recommendation list is generated, and the account managers are recommended to the customer in order from the first priority. For example, if the first priority account manager is not available to provide service, the second priority account manager is recommended to the customer.

[0102] According to the recommendation method of this application, a first recommendation list suitable for recommendation is generated based on the user profile and the account manager profile. Based on this first recommendation list, the account managers on the first recommendation list are further ranked by recommendation index to generate a second recommendation list, and the account manager ranked first on the second recommendation list is finally recommended. This method is highly compatible with customers and closely reflects their real needs. It can not only effectively improve customer satisfaction, but also enhance the marketing performance of account managers.

[0103] Figure 3 A flowchart of a method before calculating the recommendation index of each client manager in the first recommendation list using a recommendation index model according to an embodiment of the present application is schematically shown.

[0104] According to one embodiment of the present application, Figure 3 As shown, this embodiment can be used for preparation of the recommendation index model, including operations S310 to S340.

[0105] In operation S310 , a historical service data of a customer within a time period t is obtained.

[0106] Historical service data refers to the consultation, business processing, marketing recommendations, and other services a customer received from the relevant institution's business department or outlets prior to the current service transaction. This historical service data can include all historical service data received by the customer within time period t, a random set of historical service data within time period t, or the most recent historical service data. This means that this application does not specifically limit the number or time period of historical service data.

[0107] The historical service data obtained includes holographic data and voice data. Holographic data refers to the holographic data information of the entire service process recorded by the relevant institutions using the configured holographic acquisition equipment during the historical service process. Voice data refers to the voice data information of the entire service process recorded by the relevant institutions using recording equipment during the historical service process.

[0108] In operation S320, based on the a pieces of historical service data, the emotional fluctuation information of the customer in each piece of historical service data is obtained.

[0109] Furthermore, the emotion fluctuation information includes positive fluctuation, negative fluctuation and smooth fluctuation.

[0110] By using the holographic data obtained in operation S310, specific information reflecting the customer's emotions, such as the customer's expression, attitude, posture, and gesture, can be obtained, thereby analyzing the customer's emotional fluctuations.

[0111] Emotional fluctuation information may include, but is not limited to, positive fluctuation, negative fluctuation, and smooth fluctuation. It is understood that positive fluctuation is positive, indicating that the customer is gradually satisfied with the service provided by the account manager during the service, resulting in a good experience; negative fluctuation is negative, indicating that the customer is gradually disappointed with the service provided by the account manager during the service, resulting in a poor experience; and smooth fluctuation is normal, indicating that the customer is neither disappointed nor satisfied with the service provided by the account manager during the service, resulting in a neutral attitude.

[0112] In operation S330 , post-service indicator information corresponding to the customer's emotional fluctuation information is obtained.

[0113] Post-service metric information can be used as one of the data elements of the recommendation index model. Typically, after a service, customers will generate corresponding post-service metric information. For example, whether the customer purchased the recommended product, canceled the transaction, repurchased the product, or shared it on social media. Post-service metric information may include, but is not limited to, purchasing related products after being recommended, clicking on related options after being recommended, browsing related product information after being recommended, not canceling the order after purchasing the related product, repurchasing the related product, and sharing the recommended product on other platforms or with other people.

[0114] By obtaining post-service indicator information, we can indirectly express the customer's satisfaction and the customer manager's service level, which is conducive to establishing a recommendation index model.

[0115] In operation S340 , a recommendation index model is established based on the post-service indicator information and the emotion fluctuation information.

[0116] Based on the emotional fluctuation information reflected by customers on the spot read from the holographic data, as well as the post-service indicator information after being served by the account manager, a recommendation index model is established to optimize the account manager recommendation algorithm. This recommendation index model can be used to achieve effective recommendations and express the real demands of customers.

[0117] Figure 4 The flowchart of the method for obtaining emotion fluctuation information according to an embodiment of the present application is schematically shown.

[0118] According to one embodiment of the present application, Figure 4 As shown, this embodiment can obtain the customer's emotional fluctuation information in each historical service data, including operations S410 to S470.

[0119] In operation S410, holographic data and voice data of an i-th time are extracted, where i belongs to a.

[0120] The i-th representation is a service among the a historical service data obtained.

[0121] After obtaining a pieces of historical service data, the i-th historical service data is extracted, for example, the last historical service data is extracted.

[0122] In operation S420, keywords are extracted from the voice data.

[0123] Use speech recognition technology to convert voice data into text data, and then use natural language processing technology to extract keywords from the text data.

[0124] In order to divide the service phases by keywords in operation S430, keywords need to be specifically extracted to extract words related to the start or completion of the service phase. In one embodiment, the following method can be used:

[0125] 1) Obtain the account manager's service type. This can be obtained from the service type entered by the account manager before serving the customer, or from the service type selected by the customer when they wish to conduct business. Service types may include, but are not limited to, financial consulting, business processing, marketing recommendations, and other services.

[0126] 2) Different types of services may correspond to different keywords. Therefore, according to the needs, multiple related keyword libraries are established based on the service type, and the service type and the keyword library are generated and associated.

[0127] 3) Use speech recognition technology to convert speech data into text data, and then use natural language technology to extract keywords from the language data. For example, keyword extraction can be achieved through the LDA keyword extraction algorithm, the information gain keyword extraction algorithm, or other keyword extraction algorithms.

[0128] In operation S430, the i-th service is divided into a plurality of service stages based on the keywords.

[0129] Based on the keywords extracted in operation S420, the service of the i-th time is divided into stages using the service stage division model. In one embodiment, the service stages may include but are not limited to the start stage, the customer understanding stage, the recommendation stage, the customer decision stage, and the end stage.

[0130] The service stage classification model can use training samples to take keywords and service stages as input for training, and different machine learning models can be constructed for different service types. Using the service stage classification model, the time nodes at which the keywords corresponding to each service stage appear in the voice data are identified, and the time nodes at which the two keywords with the longest time span appear are used as the time nodes for that stage. For example, when words such as "consider," "discuss," and "place an order" appear in the extracted keywords, representing the end of the customer decision stage, and these three keywords appear in the voice data at 5 minutes and 6 seconds, 5 minutes and 12 seconds, and 5 minutes and 23 seconds, respectively, the end time node of the customer decision stage is the last 5 minutes and 23 seconds of the three keywords.

[0131] Furthermore, the keyword provider can be identified through holographic images based on the time the keyword appears. Since subsequent operations primarily analyze the customer's emotional characteristics, it is necessary to collect holographic images corresponding to the keyword provider. Furthermore, this allows for verification of the authenticity of the start or completion of a service phase, preventing misjudgment due to noise interference. For example, if the voice data contains other people's voices, this could result in noise.

[0132] In operation S440, a holographic image of each service stage is acquired.

[0133] Holographic data is a holographic image recorded by photoelectric image sensors such as CCD. It is converted into digital form and quantized by a data acquisition card and stored in a computer to obtain a digital holographic image.

[0134] Based on the keyword extraction in operation S420, the keyword matching algorithm can be used to obtain the time position T at which the target keyword K appears in the voice data. The time position T is used to locate the position of the key frame in the holographic data and intercept the relevant holographic image. It can be understood that the holographic image is obtained when the keyword appears in each service stage.

[0135] In operation S450, at least one emotion feature value of the customer in the holographic image is obtained by using an emotion recognition model.

[0136] Because emotional expressions are complex, a single holographic image may contain multiple emotional characteristics of a single customer. The holographic image obtained in operation S440 is input into the emotion recognition model. Analysis reveals all possible emotional characteristics expressed in the holographic image. The corresponding emotional feature value for each emotional feature is calculated to fully demonstrate the apparent expression of the customer's relevant emotional characteristic. For example, in one holographic image: Happy 0.88, Excited 0.11, Angry 0.001, Sad 0.001, Surprised 0.007, Disappointed 0.001. This indicates that the customer's primary emotion at that moment was Happy.

[0137] In operation S460 , the total emotion score of each service stage is calculated based on the emotion feature value.

[0138] Based on the emotional feature value of each holographic image obtained in operation S450, all holographic images of each service stage are processed one by one, and the emotional feature values ​​of the emotional features in the holographic images of the same stage are added together to obtain multiple emotional feature values ​​of each service stage, which represent the main emotional manifestations of the customer in the relevant stage.

[0139] Before the summation, the emotions may be divided into positive and negative attitudes according to their types, and the positive and negative attitudes may be assigned different values ​​or opposite mathematical signs (positive and negative signs).

[0140] In operation S470, the total emotion score of each service stage is analyzed to obtain the customer's emotion fluctuation information.

[0141] Emotional fluctuation information reflects customer emotions in two adjacent service stages. By comparing the total emotion scores of two adjacent service stages, the specific trend of emotional fluctuation can be determined to complete the experience fluctuation analysis.

[0142] In one embodiment, the emotion fluctuation information may include, but is not limited to, positive fluctuation, negative fluctuation, and smooth fluctuation. Positive fluctuation refers to the situation where the total emotion score of the first service stage is lower than the total emotion score of the second service stage between two adjacent service stages; negative fluctuation refers to the situation where the total emotion score of the first service stage is higher than the total emotion score of the second service stage between two adjacent service stages; and smooth fluctuation refers to the situation where the total emotion score of the first service stage is the same as the total emotion score of the second service stage between two adjacent service stages.

[0143] Figure 5 The flowchart of the method for analyzing holographic images using an emotion recognition model according to an embodiment of the present application is schematically shown.

[0144] According to one embodiment of the present application, Figure 5 As shown, at least one emotional feature value of the customer in each service stage is obtained through analysis, including operations S510 to S530.

[0145] In operation S510 , an emotion recognition model is established, which includes a plurality of emotion features.

[0146] The emotion recognition model consists of four convolutional layers, pooling layers, and fully connected layers. Each convolutional layer uses kernels of varying depths to extract features based on the number of eigenvalues.

[0147] Build and train an emotion recognition model. The model's training set samples use pictures of people with different facial expressions, postures, gestures, and attitudes. Use the training samples to train the model so that the model's predicted values ​​are closer to the true values.

[0148] The areas for emotional feature recognition include, but are not limited to, facial expressions, attitudes, postures, and gestures. Emotional features can include happiness, excitement, anger, sadness, surprise, disappointment, and other expressions derived from facial expressions. Emotional features can also include approval, opposition, acceptance, rejection, likes, dislikes, sincerity, and falsehood, derived from customer attitudes. Emotional features can also include standing upright, sitting upright, and leaning, derived from customer postures.

[0149] In operation S520, a holographic image is input to an emotion recognition model.

[0150] The holographic image is input into the trained emotion recognition model, and the convolutional neural network (CNN) is used to identify the emotional characteristics of the customer in the hologram.

[0151] In operation S530, the holographic image is analyzed, and an emotion feature value of each emotion feature is output.

[0152] Using a convolutional neural network (CNN), we can obtain the emotional feature values ​​for different emotional characteristics. These values ​​can be expressed in the form of probabilities, enabling the recognition of emotional features. For example, the facial expression feature recognition results are: happiness 0.88, excitement 0.11, anger 0.001, sadness 0.001, surprise 0.007, and disappointment 0.001.

[0153] Figure 6 The flowchart of the method for calculating the total emotion score of each service stage according to an embodiment of the present application is schematically shown.

[0154] According to one embodiment of the present application, Figure 6 As shown, the calculation method includes operations S610 to S630.

[0155] In operation S610 , each emotion feature is classified as one of a positive emotion or a negative emotion.

[0156] In one embodiment, the facial expression feature recognition results are: happiness, excitement, anger, sadness, surprise, disappointment. According to daily habits, happiness, excitement, and surprise represent positive emotions, while anger, sadness, and disappointment represent negative emotions.

[0157] In operation S620 , positive emotions are assigned positive values, and negative emotions are assigned negative values.

[0158] By distinguishing each emotion feature through operation S610, an emotion feature value with a positive or negative sign and a specific value is finally obtained. For example, the emotion feature values ​​of facial expressions are: happy +0.88, excited +0.11, angry -0.001, sad -0.001, surprised +0.007, disappointed -0.001.

[0159] In operation S630, the emotion feature values ​​of each service stage are added together.

[0160] The total emotion score of the service phase can be obtained by adding up the values ​​of each emotion feature. For example, in a service phase, the facial expression feature recognition results are: happy +0.88, excited +0.11, angry -0.001, sad -0.001, surprised +0.007, disappointed -0.001. The total emotion score S is obtained by adding up the above scores.

[0161] S=+0.88+0.11-0.001-0.001+0.007-0.001=+0.994

[0162] Similar to the above, the scores of each service stage can be obtained.

[0163] Figure 7 The flowchart of the method for analyzing emotion fluctuation information according to an embodiment of the present application is schematically shown.

[0164] According to one embodiment of the present application, Figure 7 As shown, it includes operations S710 to S720.

[0165] In operation S710 , a slope is calculated based on the total emotion scores of two adjacent service stages.

[0166] The slope λ can be obtained from the two total emotion scores. By calculating the slopes of two adjacent service stages, multiple slope values ​​can be obtained. For example, if the service is divided into five service stages using keywords, four slope values ​​can be obtained from the total emotion scores of the five service stages.

[0167] In operation S720, the slopes are added to obtain emotion fluctuation information.

[0168] The slope value represents the emotional trend of two adjacent service stages. When the slope λ>0, the customer is in a positive state and is satisfied with the service provided by the account manager at that time. When the slope λ<0, the customer is in a negative state and is disappointed with the service provided by the account manager at that time.

[0169] Sentiment fluctuation information represents the overall trend of the account manager throughout the service process, equivalent to the customer's overall experience with the account manager. Specifically, sentiment fluctuation information X can be obtained by adding up all slopes. For example, when X > 0, sentiment fluctuation information X is positive; when X < 0, sentiment fluctuation information X is negative; and when X = 0, sentiment fluctuation information X is smooth.

[0170] To further reflect customer satisfaction during the service process, the recommendation index model can be supplemented with post-service indicator information as an auxiliary recommendation indicator. Post-service indicator information includes post-service behavior information and an indicator score generated based on this post-service behavior information. For example, post-service behavior information includes: purchase, click, browse, repurchase, share, and order not cancelled. If it occurs, it is scored as 1, and if it does not occur, it is scored as 0.

[0171] During the service process, multiple post-service indicator information may be included. All indicator scores in the post-service indicator information are added together to obtain a comprehensive indicator score.

[0172] According to one embodiment of the present application, the recommendation method further includes modifying the regression model using the service evaluation feedback from customers.

[0173] Based on the service evaluation feedback from customers, when the matching degree of the first-ranked account manager is lower than the preset range, the service feature information of the account manager is re-analyzed and the first recommendation list is updated.

[0174] It can be understood that the top-ranked account manager is the one with the highest compatibility with the client. If the top-ranked account manager receives poor feedback from the client, this indicates a poor match between the manager and the client, reflecting a problem with the classification of account managers in the generated first recommendation list. Therefore, it is necessary to reanalyze the service characteristics of this account manager and use this as new sample data to replace the old sample data. The sample data for the service groups associated with the customer groups should also be updated. The updated sample data should be used for training and iterative optimization of the model.

[0175] After the service is completed, evaluation information is usually provided (for example, the customer rates or comments on the service after being served) to reflect the customer's service experience.

[0176] In one embodiment, multiple recommendation-related indicators in the recommendation index model (for example, emotional fluctuation information, post-service indicator information, evaluation information, etc.) can be calculated by weight, and weight values ​​can be assigned according to actual conditions to obtain a recommendation plan that is more suitable for the customer. The method of assigning weight values ​​will not be repeated here.

[0177] In view of the above recommendation method, this application sets up three major interactive modules according to the functions of the main operations in the recommendation method, such as Figure 8 The structure of the client manager recommendation device according to an embodiment of the present application is schematically shown.

[0178] See also Figure 8 , mainly including the branch management module, data collection module and model training module. Customers receive financial consultation, business processing, marketing recommendations, etc. from the business departments or branches of relevant institutions.

[0179] During the service, the branch staff inputs the customer's customer number and service type through the branch management module, and the recommendation device automatically recommends the best account manager for the customer.

[0180] The recommendation data for the best customer manager comes from the feedback of the model training system. The data collection module collects service data during the service process through the holographic collection equipment and voice collection equipment of the outlets, and tracks the post-service indicator information after the service is collected. This information is input into the model training module, and after data processing, model training, and model optimization, it prepares for the next recommendation.

[0181] The specific functions of the customer manager recommendation device are described in detail below.

[0182] The data acquisition module is mainly used for data collection, data acquisition and data storage. It mainly includes data acquisition module, data acquisition module and data storage module.

[0183] The data acquisition module includes a customer portrait data acquisition module, an account manager portrait data acquisition module, and a post-service indicator information acquisition module. The customer portrait data acquisition module is used to obtain customer portrait information based on the customer number; the account manager portrait data acquisition module is used to obtain account manager portrait information based on the account manager marketing code.

[0184] The data acquisition module includes a holographic data acquisition module and a voice data acquisition module. The holographic data acquisition module is used to collect the holographic data in the customer's historical services; the voice data acquisition module is used to collect the voice data in the customer's historical services.

[0185] The data storage module is mainly used for the storage of holographic data and voice data after collection, as well as the storage of customer manager recommendation lists and experience recommendation indexes for each customer during the model calculation process.

[0186] The model training module is mainly used for data processing, model training, and calculation recommendations. It mainly includes the data processing module and the list generation module.

[0187] The data processing module includes a holographic data processing module, a voice data processing module, a service stage division module, an emotion recognition module, an emotion fluctuation information analysis module, and a recommendation index calculation module. The holographic data processing module uses a CCD or other photoelectric image sensor to record holographic images, performs analog-to-digital conversion and quantization via a data acquisition card, and stores them in a computer to obtain digital holographic images. The voice data processing module uses voice recognition technology to convert speech into text and natural language processing technology to extract keywords from the text. The service stage division module uses voice data to divide financial services into stages. The emotion recognition module uses a convolutional neural network (CNN) to identify the emotional characteristics of customers in holographic images of each stage in a historical service. The emotion fluctuation information analysis module analyzes the different emotional characteristics of each stage in a historical service and calculates the total emotion score for each stage. The recommendation index calculation module uses emotion fluctuation information and post-service indicator information as training samples and uses a machine learning support vector regression (SVR) algorithm to calculate the account manager's recommendation index.

[0188] The list generation module can use machine learning clustering algorithms to build models for customer portraits and customer manager portraits, and use customer portrait features and customer manager features as training samples to obtain the first recommendation list; the list generation module can also use historical customer manager service holographic data and voice data to build a model, and sequentially call the holographic data processing module, voice data processing module, service stage division module, emotion recognition module, emotion fluctuation analysis information module, and recommendation index calculation module within the module to obtain the second customer recommendation list.

[0189] The branch management module is mainly used to generate recommendations for the best account managers after receiving customer information input, and mainly includes a service input module and an account manager recommendation module.

[0190] The service input module is mainly used by branch terminals to collect customer information and service categories from the branch management system.

[0191] The account manager recommendation module will recommend the account manager who can provide the best service for the customer to the branch terminal page.

[0192] Based on the above recommendation method, this application also provides a recommendation system. Figure 9 The system is described in detail.

[0193] Figure 9 The following schematically shows a structural block diagram of a recommendation system according to an embodiment of the present application.

[0194] like Figure 9 As shown, the recommendation system 800 of this embodiment includes a first acquisition module 810 , a second acquisition module 820 , a first generation module 830 , a calculation module 840 , a second generation module 850 and a recommendation module 860 .

[0195] Specifically, the first acquisition module 810 is used to acquire the service feature information of m account managers. In one embodiment, the first acquisition module 810 can be used to perform the operation S210 described above, which will not be repeated here.

[0196] The second acquisition module 820 is used to obtain the user portrait of the customer. In one embodiment, the second acquisition module 820 can be used to perform the operation S220 described above, which will not be repeated here.

[0197] The second acquisition module 820 is used to obtain the user portrait of the customer. In one embodiment, the second acquisition module 820 can be used to perform the operation S220 described above, which will not be repeated here.

[0198] The first generation module 830 is configured to generate a first recommendation list based on the customer's user profile and the service feature information of m account managers, where the first recommendation list includes n account managers that can be recommended to the customer, where m ≥ n, and both m and n are greater than or equal to 1. In one embodiment, the first generation module 830 can be configured to perform operation S230 described above and will not be further described here.

[0199] The calculation module 840 is used to calculate the recommendation index of each account manager in the first recommendation list using the recommendation index model. In one embodiment, the calculation module 840 can be used to perform the operation S240 described above, which will not be repeated here.

[0200] The second generation module 850 is used to sort the n account managers based on the recommendation index to generate a second recommendation list. In one embodiment, the second generation module 850 can be used to perform the operation S250 described above, which will not be repeated here.

[0201] The recommendation module 860 is used to recommend the first-ranked account manager to the client based on the second recommendation list. In one embodiment, the recommendation module 860 can be used to perform the operation S260 described above, which will not be repeated here.

[0202] According to the recommendation system of the embodiment of the present application, the recommendation method of the present application can be implemented to generate a first recommendation list suitable for recommendation based on the user profile and the account manager profile. Based on the first recommendation list, the account managers on the first recommendation list are further ranked by recommendation index to generate a second recommendation list, and the account manager ranked first on the second recommendation list is finally recommended. This method is highly compatible with customers and closely reflects their real needs. It can not only effectively improve customer satisfaction, but also enhance the marketing performance of account managers.

[0203] According to an embodiment of the present application, any multiple modules among the first acquisition module 810, the second acquisition module 820, the first generation module 830, the calculation module 840, the second generation module 850, and the recommendation module 860 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present application, at least one of the first acquisition module 810, the second acquisition module 820, the first generation module 830, the calculation module 840, the second generation module 850, and the recommendation module 860 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in an appropriate combination of any of them. Alternatively, at least one of the first acquisition module 810, the second acquisition module 820, the first generation module 830, the calculation module 840, the second generation module 850 and the recommendation module 860 can be at least partially implemented as a computer program module, which can perform the corresponding function when it is executed.

[0204] Figure 10 The block diagram schematically shows an electronic device suitable for implementing the recommendation method according to an embodiment of the present application.

[0205] like Figure 10As shown, the electronic device 900 according to an embodiment of the present application includes a processor 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage portion 908 into a random access memory (RAM) 903. The processor 901 may, for example, include a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a dedicated microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 901 may also include an onboard memory for caching purposes. The processor 901 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present application.

[0206] Various programs and data required for the operation of the electronic device 900 are stored in the RAM 903. The processor 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. The processor 901 performs various operations of the method flow according to the embodiment of the present application by executing the programs in the ROM 902 and / or the RAM 903. It should be noted that the programs may also be stored in one or more memories other than the ROM 902 and the RAM 903. The processor 901 may also perform various operations of the method flow according to the embodiment of the present application by executing the programs stored in the one or more memories.

[0207] According to an embodiment of the present application, the electronic device 900 may further include an input / output (I / O) interface 905, which is also connected to the bus 904. The electronic device 900 may further include one or more of the following components connected to the I / O interface 905: an input portion 906 including a keyboard, a mouse, etc.; an output portion 907 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage portion 908 including a hard disk; and a communication portion 909 including a network interface card such as a LAN card or a modem. The communication portion 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed in the drive 910 as needed, so that a computer program read therefrom can be installed into the storage portion 908 as needed.

[0208] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, and when the one or more programs are executed, the method according to the embodiments of this application is implemented.

[0209] According to an embodiment of the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present application, the computer-readable storage medium may include the ROM 902 and / or RAM 903 described above and / or one or more memories other than ROM 902 and RAM 903.

[0210] The present application also includes a computer program product, which includes a computer program containing program code for executing the method shown in the flowchart. When the computer program product is executed in a computer system, the program code is used to enable the computer system to implement the item recommendation method provided in the present application.

[0211] The computer program executes the above functions defined in the system / device of the embodiment of the present application when the processor 901 executes the computer program. According to the embodiment of the present application, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0212] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 909, and / or installed from a removable medium 911. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0213] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 909, and / or installed from a removable medium 911. When the computer program is executed by the processor 901, the above-mentioned functions defined in the system of the embodiment of the present application are performed. According to the embodiment of the present application, the systems, devices, means, modules, units, etc. described above can be implemented by computer program modules.

[0214] According to an embodiment of the present application, the program code for executing the computer program provided in the embodiment of the present application can be written in any combination of one or more programming languages, specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0215] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0216] Those skilled in the art will appreciate that various combinations and / or combinations of features described in the various embodiments and / or claims of this application may be made, even if such combinations or combinations are not explicitly described in this application. In particular, various combinations and / or combinations of features described in the various embodiments and / or claims of this application may be made, without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

[0217] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0218] The embodiments of the present application have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present application. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. The scope of the present application is defined by the appended claims and their equivalents. Without departing from the scope of the present application, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present application.

Claims

1. A recommendation method, characterized in that: The following steps are involved: Get the service feature information of m account managers; Obtaining a user portrait of at least one customer, wherein at least one of the customers has completed collection authorization; Generate a first recommendation list based on the user profile of the at least one customer and the service feature information of the m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer, where m≥n, and both m and n are greater than or equal to 1; Acquire a historical service data of the customer within a time period t, wherein the historical service data includes holographic data and voice data; According to the a historical service data, the emotional fluctuation information of the customer in each historical service data is obtained, wherein the method for obtaining the emotional fluctuation information of the customer comprises: extracting the holographic data and the voice data of the i-th time, wherein i belongs to a; extracting keywords from the voice data; dividing the i-th service into multiple service stages based on the keywords; obtaining a holographic image of each service stage; using an emotion recognition model, obtaining at least one emotional feature value of the customer in the holographic image; calculating the total emotional score of each service stage according to the emotional feature value; analyzing the total emotional score of each service stage to obtain the emotional fluctuation information of the customer, wherein the analyzing the total emotional score of each service stage to obtain the emotional fluctuation information of the customer comprises: calculating a slope according to the total emotional scores of two adjacent service stages; and adding the slopes to obtain the emotional fluctuation information; Obtaining post-service indicator information corresponding to the customer's emotional fluctuation information, the post-service indicator information including post-service behavior information and an indicator score generated according to the post-service behavior information; Establishing a recommendation index model based on the post-service indicator information and the emotional fluctuation information, and using the recommendation index model to calculate a recommendation index for each account manager in the first recommendation list, wherein the recommendation index is used to represent the degree of fit with the customer; Sort the n account managers based on the recommendation index to generate a second recommendation list; According to the second recommendation list, a first-ranked account manager is recommended to the client.

2. The recommendation method according to claim 1, characterized in that Obtaining at least one emotional feature value of the customer in the holographic image using an emotion recognition model includes: Build an emotion recognition model that includes multiple emotion features; Inputting holographic images into the emotion recognition model; Analyze the holographic image and output the emotional feature value of each emotional feature.

3. The recommendation method according to claim 2, characterized in that: Based on the emotional feature values, the total emotional score of each service stage is calculated, including: Classify each emotional characteristic as either positive or negative; Positive emotions are assigned positive values, and negative emotions are assigned negative values; The emotion feature values ​​of each service stage are added together.

4. The recommendation method according to claim 3, characterized in that: The emotional fluctuation information includes positive fluctuation, negative fluctuation and smooth fluctuation.

5. The recommendation method according to claim 1, characterized in that: Also includes: According to the service evaluations fed back by the customers, when the matching degree of the first-ranked account manager is lower than a preset range, the service feature information of the account manager is re-analyzed and the first recommendation list is updated.

6. The recommendation method according to claim 1, characterized in that: The service feature information includes at least one of the account manager's years of experience, resume information, marketing performance, and business scope.

7. The recommendation method according to claim 1, characterized in that: The user profile includes at least one of the customer's basic information, risk preferences, and historical purchase information.

8. A recommendation system, characterized in that include: A first acquisition module, configured to acquire service feature information of m account managers; A second acquisition module, the second acquisition module is used to obtain a user portrait of the customer; a first generating module, configured to generate a first recommendation list based on the user profile of the customer and service feature information of m account managers, wherein the first recommendation list includes n account managers that can be recommended to the customer, where m≥n, and both m and n are greater than or equal to 1; A calculation module is used to: obtain a historical service data of the customer within a time period t, wherein the historical service data includes holographic data and voice data; obtain the emotional fluctuation information of the customer in each historical service data based on the a historical service data, wherein the method for obtaining the emotional fluctuation information of the customer includes: extracting the holographic data and the voice data of the i-th time, wherein i belongs to a; extracting keywords in the voice data; based on the keywords, dividing the i-th service into multiple service stages; obtaining a holographic image of each service stage; using an emotion recognition model to obtain at least one emotional feature value of the customer in the holographic image; and calculating the emotional feature value of each service stage according to the emotional feature value. The method comprises the following steps: calculating the total emotion score of each service stage; analyzing the total emotion score of each service stage to obtain the customer's emotion fluctuation information, wherein the method comprises: calculating a slope based on the total emotion scores of two adjacent service stages; adding the slopes to obtain the emotion fluctuation information; obtaining post-service indicator information corresponding to the customer's emotion fluctuation information, wherein the post-service indicator information includes post-service behavior information and an indicator score generated based on the post-service behavior information; establishing a recommendation index model based on the post-service indicator information and the emotion fluctuation information, and using the recommendation index model to calculate the recommendation index of each customer manager in the first recommendation list; A second generating module, the second generating module being configured to: sort the n account managers based on the recommendation index to generate a second recommendation list; and A recommendation module is used to recommend a first-ranked account manager to the customer based on the second recommendation list.

9. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to perform the recommendation method according to any one of claims 1 to 7.

10. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to perform the recommendation method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the recommendation method according to any one of claims 1 to 7.

Images