Method and device for generating customer service dispatch information, electronic device, and storage medium

Abstract

The present disclosure is about a method and device, electronic device, and storage medium for generating customer service scheduling information, and relates to the field of Internet technology. The method includes: obtaining the current value of the customer service's target parameter based on the customer service's current customer service data; wherein the current customer service data includes the data of the customer service processing the customer service request at the current time, and the target parameter is used to represent the data characteristics of the current customer service data; inputting the current value of the customer service's target parameter into a customer service pressure calculation model based on machine learning to calculate the customer service's current pressure value; determining the target customer service based on the customer service's current pressure value; and allocating the current customer service request to the target customer service. The present disclosure can accurately calculate the current pressure value of the customer service, reasonably allocate the current customer service request to the target customer service based on the current pressure value, and improve the service quality of the customer service.

Description

Technical Field

[0001] The present disclosure relates to the field of Internet technology, and in particular to a method for generating customer service scheduling information, an apparatus for generating customer service scheduling information, an electronic device, and a computer-readable storage medium. Background Art

[0002] In the online customer service system, when a customer inquiry is received, the customer inquiry will generally be assigned to a target customer service representative with less customer service pressure.

[0003] In the related art, the measurement of real-time pressure is generally calculated based on the number of customers currently received by the customer service. In this way, if customer service A and customer service B both receive the same number of customers, but the questions asked by customer service A's customers are relatively simple, while the questions asked by customer service B's customers are relatively complex, if customer service A and customer service B continue to increase customers at the same rate, the pressure on customer service B will increase, resulting in uneven customer service pressure, which will lead to a decline in service quality.

[0004] In addition, we can also combine the service quality experience method, that is, combine the customer service quality (satisfaction, response time, etc.) to evaluate the customer service pressure. Low satisfaction and slow response time indicate high pressure. In this way, the weights of multiple factors can only be allocated based on experience, so the customer service pressure cannot be accurately obtained, which leads to the inability to reasonably allocate customer service.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of the present disclosure is to provide a method and device, an electronic device, and a storage medium for generating customer service scheduling information, thereby overcoming the problem of unreasonable allocation of customer service due to limitations and defects of related technologies, at least to a certain extent.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a method for generating customer service scheduling information is provided, comprising: obtaining a current value of a target parameter of the customer service based on the customer service data of the customer service; wherein the current customer service data includes data of the customer service request processed by the customer service at the current time, and the target parameter is used to represent the data feature of the current customer service data; inputting the current value of the target parameter of the customer service into a customer service pressure calculation model based on machine learning to calculate the current pressure value of the customer service; determining a target customer service based on the current pressure value of the customer service; and allocating the current customer service request to the target customer service.

[0009] In an exemplary embodiment of the present disclosure, it also includes: training the customer service pressure calculation model based on historical customer service data.

[0010] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model includes an overall pressure calculation model, and the historical customer service data includes historical customer service data of multiple customer service staff.

[0011] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model includes a pressure calculation model of a specified customer service, and the historical customer service data includes historical customer service data of the specified customer service.

[0012] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model is trained according to historical customer service data, including: determining at least one pressure factor; counting historical values ​​and historical pressure values ​​of each pressure factor within a predetermined time period according to the historical customer service data; and training the customer service pressure calculation model according to the historical values ​​of each pressure factor and the historical pressure values.

[0013] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model is trained according to the historical values ​​of each pressure factor and the historical pressure values, including: determining a target parameter from the at least one pressure factor according to the historical values ​​of each pressure factor; and training the customer service pressure calculation model according to the historical values ​​of the target parameter and the historical pressure values.

[0014] In an exemplary embodiment of the present disclosure, a target parameter is determined from the at least one pressure factor according to the historical values ​​of each pressure factor, including: calculating the correlation coefficient between every two pressure factors according to the values ​​of every two pressure factors; if the absolute value of the correlation coefficient is greater than a predetermined value, selecting one of every two pressure factors as the target parameter.

[0015] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model is a linear regression model or a piecewise linear regression model that uses different linear regression operators in different time periods.

[0016] In an exemplary embodiment of the present disclosure, it also includes: storing the current value of the customer service target parameter in the cache in the form of a key-value pair.

[0017] In an exemplary embodiment of the present disclosure, the customer service includes a first customer service and a second customer service, and the method further includes: if the current stress values ​​of the first customer service and the second customer service are the same at the same time, then obtaining the time factor offset of the first customer service and the second customer service according to the timestamps of the current customer service events of the first customer service and the second customer service respectively; and correcting the current stress values ​​of the first customer service and the second customer service respectively according to the time factor offsets of the first customer service and the second customer service.

[0018] In an exemplary embodiment of the present disclosure, the current customer service event includes any one or more of customer service login behavior, customer service logout behavior, customer service status change behavior, customer service heartbeat connection behavior, and customer service seat change behavior.

[0019] In an exemplary embodiment of the present disclosure, it also includes: determining the customer service group corresponding to each business classification according to the business classification; determining the target customer service group to which the current customer service request belongs, and determining the target customer service from the target customer service group.

[0020] According to one aspect of the present disclosure, there is provided a device for generating customer service scheduling information, including: a current value acquisition module, used to calculate the current value of a target parameter of a customer service according to the current customer service data of the customer service; wherein the current customer service data includes data of a customer service request processed by the customer service at the current time, and the target parameter is used to represent the data feature of the current customer service data; a pressure value calculation module, used to input the current value of the target parameter of the customer service into a customer service pressure calculation model based on machine learning, and calculate the current pressure value of the customer service; a target customer service determination module, used to determine a target customer service according to the current pressure value of the customer service; and a customer service request allocation module, used to allocate the current customer service request to the target customer service.

[0021] According to one aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-described methods for generating customer service dispatch information by executing the executable instructions.

[0022] According to one aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for generating customer service scheduling information described in any one of the above is implemented.

[0023] In a method for generating customer service scheduling information, an apparatus for generating customer service scheduling information, an electronic device, and a computer-readable storage medium provided in an exemplary embodiment of the present disclosure, on the one hand, the current value of the target parameter is obtained through the current customer service data of the customer service, and the current stress value of the customer service is calculated based on the current value of the target parameter and a trained customer service stress calculation model, so that a more accurate current stress value can be obtained through target parameters of multiple dimensions; on the other hand, because the obtained current stress value is more accurate, the target customer service can be reasonably determined according to the accurate current stress value, and then the current customer service requests can be reasonably allocated, thereby improving the customer service quality.

[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0026] Figure 1 A schematic diagram of a system architecture for implementing a method for generating customer service scheduling information in an exemplary embodiment of the present disclosure is schematically shown;

[0027] Figure 2 A schematic diagram schematically illustrates a method for generating customer service scheduling information in an exemplary embodiment of the present disclosure;

[0028] Figure 3 Schematically showing a flow chart of training a customer service stress calculation model in an exemplary embodiment of the present disclosure;

[0029] Figure 4 The process of obtaining target parameters in an exemplary embodiment of the present disclosure is schematically shown;

[0030] Figure 5 Schematically illustrates a storage structure of current values ​​of target parameters in an exemplary embodiment of the present disclosure;

[0031] Figure 6 The model training flow chart in the exemplary embodiment of the present disclosure is schematically shown;

[0032] Figure 7 A block diagram schematically illustrates a device for generating customer service scheduling information in an exemplary embodiment of the present disclosure;

[0033] Figure 8The structure of a computer system of an electronic device in an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0034] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0035] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0036] In this example implementation, a system architecture for implementing a customer service dynamic scheduling method is first provided. Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0037] The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send request instructions, etc. Various communication client applications can be installed on the terminal devices 101, 102, 103, such as image processing applications, shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0038] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0039] The server 105 may be a server that provides various services, such as a backend management server (only an example) that provides support for shopping websites browsed by users using the terminal devices 101, 102, and 103. The backend management server may analyze and process the received data such as product information query requests, and feed back the processing results (such as target push information, product information - only an example) to the terminal device.

[0040] It should be noted that the method for generating customer service scheduling information provided in the embodiment of the present application is generally executed by the server 105 , and accordingly, the device for generating customer service scheduling information is generally set in the terminal device 101 .

[0041] based on Figure 1 In the system architecture, this example implementation provides a method for generating customer service scheduling information, which can be applied to any scenario of processing customer service requests. Figure 2 As shown, the method for generating customer service scheduling information in this exemplary embodiment is specifically described.

[0042] In step S210, the current value of the customer service target parameter is obtained based on the customer service data of the customer service; wherein the current customer service data includes data of the customer service request processed by the customer service at the current time, and the target parameter is used to represent the data characteristics of the current customer service data.

[0043] In this exemplary embodiment, the current customer service data refers to the customer service data corresponding to the customer service request being processed by a certain customer service at the current time, and the customer service data includes but is not limited to all data of the customer service request processed by the customer service, such as the number of reply messages, reply duration, response time, etc. The target parameter refers to a plurality of parameters used to describe the data characteristics of the current customer service data of the customer service. Through multiple target parameters, the current customer service data can be analyzed from multiple dimensions, making the data more comprehensive and accurate. The current value of the target parameter refers to the actual value of each target parameter at the current time.

[0044] Next, in step S220, the current value of the customer service target parameter is input into the customer service pressure calculation model based on machine learning to calculate the current pressure value of the customer service.

[0045] In this exemplary embodiment, after obtaining the current value of the target parameter of the customer service, the current value can be input as training data into the customer service pressure calculation model based on machine learning, so as to calculate the current pressure value of the customer service through the trained customer service pressure calculation model based on machine learning. It should be noted that in order to accurately calculate the current pressure value of the customer service, it is first necessary to train the customer service pressure calculation model to obtain a model with better performance. Specifically, the customer service pressure calculation model can be trained based on historical customer service data. Historical customer service data refers to customer service data within a predetermined time period. In order to ensure the validity of the results, the predetermined time period can be set to a time period close to the current time, such as within one month or within three months, etc.

[0046] It should be noted that, for the case where there are a large number of customer service requests, in order to improve user satisfaction and reasonably allocate customer service, it is also necessary to accurately calculate the current pressure value of the customer service, and allocate the target customer service for the current customer service request according to the calculated current pressure value. In addition, it is found in practice that for the case where there are a large number of customer service requests, the customer service pressure calculation model obtained by training the historical customer service data after denoising or other smoothing processing is similar to the customer service pressure calculation model obtained by training the historical customer service data without denoising and smoothing processing. Therefore, in order to reduce the operation steps, the method in this exemplary embodiment can be directly adopted to train the customer service pressure calculation model through the historical customer service data of a predetermined time period, so that the current pressure value of the customer service can be calculated by the customer service pressure training model in any case, and the application range is wider. Furthermore, in order to meet the needs of the case where there are a large number of customer service requests, a preset number of temporary customer services with initial pressure values ​​can be added, and the initial pressure value is obtained by the customer service pressure calculation model trained according to the historical customer service data of multiple customers in a predetermined time period. Then, the current value of the target parameter of the temporary customer service can be obtained, and it can be input into the trained customer service pressure calculation model to calculate the current pressure value of the temporary customer service. In this way, the quality of the customer service request processing by the customer service can be reflected according to the current pressure value.

[0047] When the customer service pressure calculation model is trained according to the historical customer service data, the following two methods may be included: Method 1, the customer service pressure calculation model includes an overall pressure calculation model, and the historical customer service data includes the historical customer service data of multiple customer services. The overall pressure calculation model refers to a pressure calculation model for multiple customer services. In this case, the historical customer service data used to train the overall pressure calculation model may also include the historical customer service data of multiple customer services. The multiple customer services may be all customer services, or may be some of the customer services among all customer services, for example, the historical customer service data of all customer services in the customer service group to which the customer service belongs. Through method 1, the overall pressure calculation model is trained based on the historical customer service data of multiple customer services, and a trained overall pressure calculation model can be obtained. If the customer service belongs to a new customer service with less or no historical data, the trained overall pressure calculation model can be used to calculate the pressure of the new customer service in real-time calculation. For example, customer service F belongs to a new customer service, and there will not be much historical customer service data. At this time, an overall pressure calculation model can be trained by the historical customer service data of all customer services belonging to the same customer service group as customer service F, so as to calculate the pressure of customer service F according to the trained overall pressure calculation model. In this way, when there is less historical customer service data, the customer service pressure can be calculated through the overall pressure calculation model, avoiding the problem of being unable to calculate the pressure or not being able to accurately calculate the pressure due to the absence of historical customer service data.

[0048] Method 2: The customer service pressure calculation model includes a pressure calculation model of a designated customer service, and the historical customer service data includes the historical customer service data of the designated customer service. The pressure calculation model of the designated customer service refers to a pressure calculation model for each customer service. In this case, the historical customer service data used to train the pressure calculation model may also include the historical customer service data of the designated customer service, that is, the historical customer service data of each customer service. Based on method 2, the pressure calculation model of the designated customer service is trained by the historical customer service data of the designated customer service, and a trained pressure calculation model of the designated customer service can be obtained. For example, if a customer service belongs to a customer service with a lot of historical data, the pressure calculation model of the designated customer service can be used to calculate the pressure of the designated customer service in real-time calculation. For example, if customer service B belongs to a customer service with a lot of historical customer service data, the pressure calculation model for customer service B is trained by the historical customer service data of customer service B, so that the pressure of customer service B is calculated according to the trained pressure calculation model for customer service B. In this way, when there is a lot of historical data of the customer service, the pressure of each customer service can be accurately calculated by the pressure calculation model for the customer service.

[0049] Figure 3 The process of training the customer service pressure calculation model is schematically shown in FIG. , which specifically includes the following steps:

[0050] In step S310 , at least one pressure factor is determined.

[0051] In this exemplary embodiment, the dimensions describing customer service pressure can reach dozens of indicators. Here, by filtering dozens of indicators, at least one pressure factor with better effect can be selected from them, so as to accurately describe the pressure from multiple dimensions through at least one pressure factor. The at least one pressure factor includes any one or more of the average number of messages in the conversation, the number of messages typed by the customer service in the conversation, the first reply time of the customer service, the average response time of the conversation, the average number of customer messages, the number of emoticons, and the number of conversations. Among them, the average number of messages in the conversation refers to the number of all messages in the current conversation between the customer and the customer service. The number of messages typed by the customer service in the conversation refers to the data that the customer service is purely typed by hand in this conversation, excluding copy and paste, emoticons, etc. The first reply response time of the customer service refers to the period from the customer initiating the consultation to the first reply of the customer service in a conversation. The average response time of the conversation refers to the period from the customer sending a sentence to the first message replied by the customer service. The average number of customer messages refers to the average number of messages sent by the customer in each conversation in N conversations. The number of emoticons refers to the average number of emoticons sent by the customer service in N conversations. The number of conversations refers to the average number of conversations received by each customer service per day for N customer service staff. After determining at least one stress factor, each stress factor can be expressed as i To represent, thereby forming a set S representing at least one pressure factor.

[0052] In step S320, historical values ​​of various pressure factors and historical pressure values ​​within a predetermined time period are counted based on the historical customer service data.

[0053] In this exemplary embodiment, based on the set S formed in step S310, the number of customers (pressure) received by the customer service staff at the same time can be determined as the dependent variable Y. After obtaining the historical sums and historical pressure values ​​of each pressure factor, several (X i ,Y) value pair.

[0054] Furthermore, in step S330, the customer service pressure calculation model is trained according to the historical values ​​of each pressure factor and the historical pressure values.

[0055] In this exemplary embodiment, the number of (X) obtained in step S320 can be i ,Y) value pairs are used as training data to train the customer service pressure calculation model, and a trained customer service pressure calculation model is obtained. Figure 3As shown in , the specific process of training the customer service pressure calculation model according to the historical values ​​of each pressure factor and the historical pressure values ​​described in step S330 includes the following steps: Step S331, determining the target parameter from the at least one pressure factor according to the historical values ​​of each pressure factor. Since there may be factors describing the same property among multiple pressure factors, in order to avoid repeated calculations, at least one pressure factor may be screened to obtain a valid target parameter. The target parameter is a subset of at least one pressure factor, which can be represented by S', which may include some or all of the pressure factors determined in step S310. Step S330 also includes step S332, training the customer service pressure calculation model according to the historical values ​​of the target parameter and the historical pressure values.

[0056] The specific step of determining the target parameter from the at least one pressure factor according to the historical values ​​of each pressure factor in step S331 includes: step S3311, calculating the correlation coefficient between each two pressure factors according to the values ​​of each two pressure factors. Specifically, the correlation coefficient between each two pressure factors in at least one pressure factor can be calculated to perform correlation analysis. The correlation coefficient between each two pressure factors can be calculated according to formula (1):

[0057]

[0058] Among them, Cov(X i ,X i+1 ) is X i and X i+1 The covariance of Var[X i ] is X i The variance of Var[X i+1 ] is X i+1 The variance of .

[0059] Since |r(X i ,X i+1 )| has a value range of (0, 1), so |r(X i ,X i+1 )|The closer it is to 1, the more it proves that the two factors X i and X i+1 The stronger the correlation, the more likely the two factors are describing the same thing.

[0060] Step S3312, if the absolute value of the correlation coefficient is greater than a predetermined value, one of the two stress factors is selected as the target parameter. The predetermined value can be set according to the actual situation, for example, set to 0.8, 0.9, etc., and the predetermined value of 0.8 is used as an example for explanation. If the correlation between the two stress factors is relatively large, it means that the two stress factors describe the same thing. Only by discarding one of the two stress factors with a strong correlation can the influencing factors of one thing be correctly described, while reducing the amount of data calculation. For example, if the correlation coefficient between the average number of customer messages and the number of conversations is greater than 0.8, either the average number of customer messages or the number of conversations can be discarded. When discarding a stress factor, one stress factor can be randomly discarded. And the remaining stress factors are used as target parameters. Through the above formula (1), the correlation coefficient between every two stress factors can be calculated. By using the correlation coefficient between the stress factors, the set S' of stress factors with little correlation, i.e., the target parameter, can be obtained, where S' is a subset of the set S representing at least one stress factor.

[0061] In addition, when the dimensions of the pressure factors are controllable, the PCA (Principal Component Analysis) method can also be used to reduce the dimensions of all pressure factors to obtain the target parameters.

[0062] Figure 4 The whole process of obtaining the target parameters is shown in FIG. The first step is to define the pressure factor; the second step is to perform offline statistics on all pressure factors to obtain the original pressure factor set S; the third step is to perform correlation analysis on every two pressure factors in the original pressure factor set S; the fourth step is to remove one of the two pressure factors with greater correlation to obtain the target parameter, that is, the effective pressure factor set S'. Figure 4 On this basis, a more accurate model can be obtained by training the pressure calculation model through the historical values ​​of the target parameters and the historical pressure values.

[0063] After obtaining the target parameters, for each customer service that is working, when any customer service event arrives, the target parameter set S'={X i}, but when the customer service event arrives, only the online real-time data of the target parameter is known, but the customer service pressure Y is unknown. At this time, the current value of the customer service target parameter can be stored in the cache in the form of a key-value pair. Figure 5 As shown in , the current value of the customer service target parameter is stored in the customer service account as the key, and the specific factor X iThe English name of the redis cache structure is field. Since it is stored in the form of key-value pairs, when you need to obtain the current value of the customer service target parameter, you can get all the values ​​at once according to the customer service account, which improves data processing efficiency. For example, the pin of the customer service account is zhangsan, which represents the target parameter X of the average response time of the customer service. 1 =8, representing the target parameter X for the average number of sessions 2 =13, representing the target parameter X of the proportion of customer service manual messages 3 =0.68 and so on.

[0064] Next, combine Figure 6 The model training process is illustrated in FIG. 1. When the historical values ​​of the target parameters of the customer service are obtained by offline calculation, the value pair (X i ,Y). Then the value pair can be input into the customer service pressure calculation model to train the customer service pressure calculation model. The customer service pressure calculation model can be a linear regression model or a segmented linear regression model that uses different linear regression operators for different time periods. Specifically, when the customer service pressure calculation model is a linear regression model, its expression can be as shown in formula (2):

[0065]

[0066] The process of training the customer service pressure calculation model can be understood as the process of obtaining the optimal coefficient. Therefore, after training, the factor coefficient μ can be calculated. i and intercept b, thereby obtaining a trained customer service pressure calculation model corresponding to the factor coefficient and intercept value.

[0067] On this basis, the segmented linear regression algorithm can also be used to train the model according to the business characteristics. For example, the business at different times of the day is different, and the customer service personnel arranged are also different. Different linear regression operators can be used for different time periods, that is, the customer service pressure calculation model is a segmented linear regression model that uses different linear regression operators for different time periods. In this case, the model can be expressed by formula (3):

[0068]

[0069] The process of training the customer service pressure calculation model here can be understood as the process of obtaining the optimal coefficient. Therefore, after the training, the factor coefficient and the intercept can be calculated to obtain the trained customer service pressure calculation model corresponding to the factor coefficient and the intercept value.

[0070] It should be noted that the customer service pressure calculation model can also be other suitable machine learning models, including but not limited to regression classification, decision number, support vector machine, Bayesian algorithm, neural network and other methods and their variants.

[0071] After obtaining the current value of the customer service target parameter and the trained customer service pressure calculation model, the current value of the customer service target parameter can be input into the trained customer service pressure calculation model, and the current pressure value of the customer service can be calculated in real time when the current customer service event arrives. Among them, the current customer service event includes any one or more of the customer service login behavior, customer service logout behavior, customer service status change behavior, customer service heartbeat connection behavior, and customer service seat change behavior. Among them, customer service status change behaviors generally include hanging up, online, leaving, etc. The customer service heartbeat connection behavior is to tolerate physical factors such as network failures and delays. Generally, the client will periodically send survival detection data to the server to determine whether the client is in the network and improve the reliability of message sending in later chats. It is a standard fault-tolerant mechanism in the industry.

[0072] It should be added that if the current pressure values ​​of multiple customer services are equal through the trained customer service pressure calculation model, it is necessary to correct the current pressure value of each customer service with the same current pressure value, so as to select a more suitable customer service. For example, the current pressure values ​​of the first customer service and the second customer service are equal. It specifically includes two steps: Step 1, if the current pressure values ​​of the first customer service and the second customer service are the same at the same time, the time factor offset of the first customer service and the second customer service is obtained according to the timestamp of the current customer service event of the first customer service and the second customer service respectively. The timestamp can be represented by current_time, which refers to the time when the current customer service event arrives at the first customer service and the second customer service. The time factor offset refers to the value obtained by combining the timestamp with a reference value to represent the time difference, which can be represented by offSet. Among them, the reference value can be the maximum value of a long integer number, and offSet is only a decimal between [0,1), but the sensitivity is very high.

[0073] Step 2: Correct the current pressure values ​​of the first customer service and the second customer service respectively according to the time factor offset of the first customer service and the second customer service. In order to avoid the current pressure values ​​of multiple customer services being the same, the current pressure values ​​of the first customer service and the second customer service may be corrected respectively by the time factor offset to distinguish the current pressure values ​​of multiple customer services. The specific correction process is shown in formula (4):

[0074] offSet=current_time / Long.MAX_VALUE (4)

[0075] For example, the specific value of the maximum value of the long integer Long.MAX_VALUE is 9223372036854775807. If the current pressure value of customer service A and customer service B is the same, and the arrival time of the current customer service event of customer service A is 2018-08-22 09:00:00, it is converted into Unix millisecond timestamp 1534899600000, so the time factor offset of customer service A is 1534899600000 / 9223372036854775807. Similarly, if the arrival time of the current customer service event of customer service B is 2018-08-22 09:00:01, the millisecond timestamp converted into Unix is ​​1534899601000, and the offset of customer service B = 1534899601000 / 9223372036854775807.

[0076] If the current pressure values ​​of more than three customer service staff are the same, the current pressure value of each customer service staff can be corrected according to the above formula (4), which is not described in detail here.

[0077] By correcting the current pressure value, a corrected current pressure value can be obtained. The corrected current pressure value can be obtained by performing a logical operation on the current pressure value calculated by the trained customer service pressure calculation model and the time factor offset offSet. For example, the two can be added or subtracted, etc. Here, the addition operation is used as an example for explanation. The corrected current pressure value can be the sum of the current pressure value calculated by the trained customer service pressure calculation model and the time factor offset offSet.

[0078] In this example, since the customer service pressure calculation model is trained by target parameters of multiple dimensions in at least one pressure factor, the customer service pressure calculation model can be made more accurate and have better performance. In addition, by calculating the current pressure value of each customer service through the trained customer service pressure calculation model and correcting the current pressure value, a more accurate pressure value can be obtained.

[0079] In step S230, the target customer service is determined according to the current stress value of the customer service.

[0080] In this exemplary embodiment, on the basis of step S220, a target customer service can be determined from multiple customer services according to the current pressure value of the customer service obtained by the trained customer service pressure calculation model or the corrected current pressure value. Specifically, when determining the target customer service, the customer service group corresponding to each business classification can be determined according to the business classification; the target customer service group to which the current customer service request belongs is determined, and the target customer service is determined from the target customer service group. Business classification refers to a customer service group that receives a specific business, and a customer service group includes multiple online customer services. For example, the customer service group includes a pre-sales skill group responsible for pre-sales customer consultation reception, a refrigerator skill group responsible for customer consultation reception for refrigerator business, and a mobile phone skill group responsible for customer consultation reception for mobile phone business.

[0081] The current customer service request may be a customer service request sent by a customer through a chat software or a chat page in a smart phone or other smart terminal, and may be sent in text or voice. Next, the skill group to which the current customer service request belongs may be identified through the interface to which the current customer service request is sent, or the skill group to which the current customer service request belongs may be determined through the customer's account information or other purchase information.

[0082] After determining which skill group the current customer service request belongs to, multiple customer services in the skill group can be screened to obtain the customer service with the smallest current stress value as the target customer service. Specifically, multiple customer services refer to the online customer services in the skill group. The current stress values ​​of all online customer services can be arranged in order from small to large, and the customer service with the smallest current stress value ranked first can be used as the target customer service for the current customer service request. By using the customer service with the smallest current stress value as the target customer service, the problem of different customer services adding customers at the same rate when the difficulty of the problem is different, which makes the pressure of some customer services increase, is avoided, and the pressure of customer services is balanced, thereby improving the service quality.

[0083] In step S240, the current customer service request is assigned to the target customer service.

[0084] In this step, based on step S230, the current customer service request can be assigned to the target customer service with the smallest current pressure value, so that the target customer service can quickly process the current customer service request, provide services to customers in a timely manner, and improve user experience.

[0085] For example, if the current customer service request is sent by the user by clicking on the customer service chat page of the mobile page, it can be considered that the current customer service request belongs to the mobile skill group. Among all the customer services in the mobile skill group, the current pressure values ​​of the customer services calculated by the customer service pressure calculation model are: customer service A is 50, customer service B is 50, customer service C is 60, and customer service D is 80. Since the current pressure values ​​of customer service A and customer service B are the same, the time factor offset of customer service A and customer service B is obtained according to the timestamp to correct the current pressure values ​​of customer service A and customer service B. If after correction, customer service A is 50.05, customer service B is 50.03, customer service C is 60, and customer service D is 80. Then multiple customer services can be arranged in the queue in the order of the current pressure value from small to large. Through steps S210 to S240, when the current customer service request sent by the customer is received, the target customer service can be directly taken out from the head of the mobile skill group corresponding to the current customer service request for allocation, that is, the target customer service allocated for the current customer service request is customer service B, so that customer service B can serve the customer in time. In this way, compared with the empirical method, the weights of the target parameters can be accurately determined, so that the customer service pressure can be accurately obtained, and then the target customer service can be efficiently and reasonably allocated to the current customer service request according to the customer service pressure, thereby reducing customer waiting time and improving customer service quality.

[0086] The present disclosure also provides a device for generating customer service scheduling information. Figure 7 As shown, the device 700 for generating customer service scheduling information may include:

[0087] The current value acquisition module 701 is used to calculate the current value of the target parameter of the customer service according to the current customer service data of the customer service; wherein the current customer service data includes the data of the customer service request processed by the customer service at the current time, and the target parameter is used to represent the data characteristics of the current customer service data;

[0088] A pressure value calculation module 702 is used to input the current value of the customer service target parameter into the customer service pressure calculation model based on machine learning to calculate the current pressure value of the customer service;

[0089] The target customer service determination module 703 is used to determine the target customer service according to the current stress value of the customer service;

[0090] The customer service request allocation module 704 is used to allocate the current customer service request to the target customer service.

[0091] In an exemplary embodiment of the present disclosure, it also includes: a model training module, which is used to train the customer service pressure calculation model based on historical customer service data.

[0092] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model includes an overall pressure calculation model, and the historical customer service data includes historical customer service data of multiple customer service staff.

[0093] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model includes a pressure calculation model of a specified customer service, and the historical customer service data includes historical customer service data of the specified customer service.

[0094] In an exemplary embodiment of the present disclosure, the model training module includes: a stress factor determination module, used to determine at least one stress factor; a historical value statistics module, used to count the historical values ​​and historical stress values ​​of each stress factor within a predetermined time period based on the historical customer service data; and a training control module, used to train the customer service stress calculation model based on the historical values ​​of each stress factor and the historical stress values.

[0095] In an exemplary embodiment of the present disclosure, the training control module includes: a target parameter selection module, used to determine the target parameter from the at least one stress factor according to the historical values ​​of each stress factor; a training module, used to train the customer service pressure calculation model according to the historical values ​​of the target parameter and the historical pressure values.

[0096] In an exemplary embodiment of the present disclosure, the target parameter selection module includes: a correlation coefficient calculation module, which is used to calculate the correlation coefficient between every two pressure factors according to the values ​​of every two pressure factors; and a target parameter determination module, which is used to select one of every two pressure factors as the target parameter if the absolute value of the correlation coefficient is greater than a predetermined value.

[0097] In an exemplary embodiment of the present disclosure, the customer service pressure calculation model is a linear regression model or a piecewise linear regression model that uses different linear regression operators in different time periods.

[0098] In an exemplary embodiment of the present disclosure, it also includes: a current value storage module, which is used to store the current value of the customer service target parameter in the cache in the form of a key-value pair.

[0099] In an exemplary embodiment of the present disclosure, the customer service includes a first customer service and a second customer service, and the device also includes: an offset calculation module, which is used to obtain the time factor offset of the first customer service and the second customer service according to the timestamp of the current customer service events of the first customer service and the second customer service respectively if the current pressure values ​​of the first customer service and the second customer service are the same at the same time; and a correction module, which is used to correct the current pressure values ​​of the first customer service and the second customer service respectively according to the time factor offset of the first customer service and the second customer service.

[0100] In an exemplary embodiment of the present disclosure, the current customer service event includes any one or more of a customer service login behavior, a customer service logout behavior, a customer service status change behavior, a customer service heartbeat connection behavior, and a customer service seat change behavior.

[0101] In an exemplary embodiment of the present disclosure, it further includes: a customer service group determination module, configured to determine the customer service groups corresponding to each business classification according to the business classification; a target determination module, configured to determine the target customer service group to which the current customer service request belongs, and determine the target customer service from the target customer service group.

[0102] It should be noted that the specific details of each module in the above device for generating customer service scheduling information have been described in detail in the corresponding method for generating customer service scheduling information, and thus will not be elaborated herein.

[0103] Figure 8 The computer system 800 of the illustrated electronic device is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present invention.

[0104] As Figure 8 shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 802 or the program loaded from the storage section 208 into the random access memory (RAM) 803. In the RAM 803, various programs and data required for system operation are also stored. The CPU 801, ROM 802, and RAM 803 are connected to each other via a bus 804. The input / output (I / O) interface 805 is also connected to the bus 804.

[0105] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that the computer program read from it can be installed into the storage section 808 as needed.

[0106] In particular, according to an embodiment of the present invention, the process described below with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 809, and / or installed from a removable medium 811. When the computer program is executed by a central processing unit (CPU) 801, various functions defined in the system of the present application are executed.

[0107] It should be noted that the computer-readable storage medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, 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, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0108] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a 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 a different order from the order 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 flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0109] The units involved in the embodiments of the present invention may be implemented by software or hardware, and the units described may also be arranged in a processor. The names of these units do not, in some cases, limit the units themselves.

[0110] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the following embodiment. For example, the electronic device may implement the following Figure 2 The steps shown.

[0111] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A method for generating customer service scheduling information, It is characterized in that include: Obtaining a current value of a target parameter of the customer service according to the current customer service data of the customer service; wherein the current customer service data includes data of the customer service request processed by the customer service at the current time, and the target parameter is used to represent the data characteristics of the current customer service data; Inputting the current value of the target parameter of the customer service into the customer service pressure calculation model based on machine learning to calculate the current pressure value of the customer service; the customer service includes a first customer service and a second customer service; If the current pressure values ​​of the first customer service and the second customer service are the same at the same time, the time factor offsets of the first customer service and the second customer service are obtained respectively according to the ratios of the timestamps of the current customer service events of the first customer service and the second customer service to the reference values; the current customer service events include any one or more of customer service login behavior, customer service logout behavior, customer service status change behavior, customer service heartbeat connection behavior, and customer service seat change behavior; Correcting the current pressure values ​​of the first customer service and the second customer service respectively according to the time factor offset of the first customer service and the second customer service, wherein the corrected current pressure value is the sum of the current pressure value and the time factor offset; Determine the target customer service according to the customer service's current stress value; Assign the current customer service request to the target customer service.

2. The method according to claim 1, It is characterized in that Also includes: The customer service pressure calculation model is trained based on historical customer service data.

3. The method according to claim 2, It is characterized in that The customer service pressure calculation model includes an overall pressure calculation model, and the historical customer service data includes historical customer service data of multiple customer services.

4. The method according to claim 2, It is characterized in that The customer service pressure calculation model includes a pressure calculation model of a specified customer service, and the historical customer service data includes historical customer service data of the specified customer service.

5. The method according to claim 2, It is characterized in that The customer service pressure calculation model is trained based on historical customer service data, including: identifying at least one stress factor; According to the historical customer service data, historical values ​​of various pressure factors and historical pressure values ​​within a predetermined time period are counted; The customer service pressure calculation model is trained according to the historical values ​​of each pressure factor and the historical pressure values.

6. The method according to claim 5, It is characterized in that Training the customer service pressure calculation model according to the historical values ​​of each pressure factor and the historical pressure value includes: determining a target parameter from the at least one stress factor based on historical values ​​of the respective stress factors; The customer service pressure calculation model is trained according to the historical value of the target parameter and the historical pressure value.

7. The method according to claim 6, It is characterized in that Determining a target parameter from the at least one pressure factor according to historical values ​​of each pressure factor includes: Calculate the correlation coefficient between each two stress factors according to their values; If the absolute value of the correlation coefficient is greater than a predetermined value, one of every two pressure factors is selected as the target parameter.

8. The method according to claim 6, It is characterized in that The customer service pressure calculation model is a linear regression model or a piecewise linear regression model that uses different linear regression operators in different time periods.

9. The method according to claim 1, It is characterized in that Also includes: Store the current value of the customer service target parameter in the cache as a key-value pair.

10. The method according to claim 1, It is characterized in that Also includes: According to the business classification, determine the customer service group corresponding to each business classification; Determine a target customer service group to which the current customer service request belongs, and determine the target customer service from the target customer service group.

11. A device for generating customer service dispatch information, It is characterized in that include: A current value acquisition module, used to calculate the current value of the target parameter of the customer service according to the current customer service data of the customer service; wherein the current customer service data includes the data of the customer service request processed by the customer service at the current time, and the target parameter is used to represent the data characteristics of the current customer service data; A pressure value calculation module, used to input the current value of the target parameter of the customer service into the customer service pressure calculation model based on machine learning to calculate the current pressure value of the customer service; the customer service includes a first customer service and a second customer service; An offset calculation module is used to obtain the time factor offset of the first customer service and the second customer service respectively according to the ratio of the timestamp of the current customer service event of the first customer service and the second customer service to a reference value if the current pressure values ​​of the first customer service and the second customer service are the same at the same time; the current customer service event includes any one or more of customer service login behavior, customer service logout behavior, customer service status change behavior, customer service heartbeat connection behavior, and customer service seat change behavior; A correction module, used to correct the current pressure values ​​of the first customer service and the second customer service respectively according to the time factor offset of the first customer service and the second customer service, and the corrected current pressure value is the sum of the current pressure value and the time factor offset; A target customer service determination module is used to determine the target customer service according to the current pressure value of the customer service; The customer service request allocation module is used to allocate the current customer service request to the target customer service.

12. An electronic device, It is characterized in that include: processor; as well as A memory, configured to store executable instructions of the processor; Wherein, the processor is configured to execute the method for generating customer service scheduling information as described in any one of claims 1-10 by executing the executable instructions.

13. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method for generating customer service scheduling information described in any one of claims 1 to 10 is implemented.

Images