Seat grouping method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a seat grouping method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: receiving an agent grouping request, wherein the agent grouping request carries an agent to be grouped; based on a pre-trained seat capability evaluation model, processing seat data corresponding to the to-be-grouped seats to obtain index feature vectors; and based on a seat grouping model obtained by training, processing the index feature vector to obtain a grouping result corresponding to each seat to be grouped. By adopting the method, the grouping efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a seat grouping method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the growth of customer service demand and the improvement of service quality requirements in business operations, the traditional manual agent allocation method can no longer meet the needs of efficient and accurate customer service. The customer service agent allocation in traditional technology usually relies on manual experience or simple matching rules, which is inefficient and difficult to adapt to the ever-changing customer needs and service environment. Summary of the invention

[0003] Based on this, it is necessary to provide a seat grouping method, device, computer equipment, computer-readable storage medium and computer program product that can improve processing efficiency in response to the above technical problems.

[0004] In a first aspect, the present application provides a seat grouping method, the method comprising:

[0005] receiving a seat grouping request, wherein the seat grouping request carries seats to be grouped;

[0006] Based on the pre-trained seat capability evaluation model, the seat data corresponding to each of the seats to be grouped are processed to obtain an indicator feature vector;

[0007] Based on the seat grouping model obtained through training, the indicator feature vector is processed to obtain the grouping result corresponding to each of the seats to be grouped.

[0008] In one embodiment, the grouping result includes the grouping probability of the seats to be grouped belonging to each group; after the seat grouping model obtained through training processes the indicator feature vector to obtain the grouping result corresponding to each seat to be grouped, the method includes:

[0009] Determine the initial group corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group;

[0010] Determine the grouping probability corresponding to the seats to be grouped that are in the same initial group;

[0011] In the case where the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial grouping, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial grouping from the initial grouping;

[0012] When the grouping probability of the seats to be grouped meets the grouping threshold of the initial grouping, a grouping label corresponding to the initial grouping is generated for the seats to be grouped.

[0013] In one embodiment, before determining the initial group corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group, the method further includes:

[0014] Determine whether each of the seats to be grouped is on duty;

[0015] If the seat to be grouped is not on duty, deleting the seat to be grouped;

[0016] In the case that the seat to be grouped is on duty, the step of determining the initial group corresponding to the seat to be grouped based on the grouping probability that the seat to be grouped belongs to each group is continued.

[0017] In one embodiment, the training method of the agent capability evaluation model and the agent grouping model includes:

[0018] Acquire target sample data, and extract key features from the target sample data;

[0019] Obtaining a weight assigned to each of the key features;

[0020] Based on the weights corresponding to the key features and the scores of the key features corresponding to the target sample data, a sample index feature vector corresponding to each target sample data is obtained;

[0021] Based on the sample indicator feature vector, the preset logistic regression model and the total number of seat groups, a corresponding training grouping result is obtained;

[0022] Based on the training grouping results and the target grouping results corresponding to the target sample data, training the weights assigned to the key features and / or the model parameters in the preset logistic regression model;

[0023] Obtaining an agent capability assessment model based on the weights assigned to each of the key features obtained through training;

[0024] The seat grouping model is obtained based on the model parameters in the preset logistic regression model obtained through training.

[0025] In one embodiment, obtaining target sample data includes:

[0026] Obtain historical agent business data;

[0027] Cleaning the historical agent business data;

[0028] Performing missing data detection and abnormal data detection on the cleaned historical agent business data;

[0029] The business data with missing data or abnormal data in the historical agent business data is processed to obtain target sample data.

[0030] In one embodiment, the method further comprises:

[0031] Obtaining an agent evaluation form, and obtaining satisfaction of each group based on the agent evaluation form;

[0032] In the case where the satisfaction level of the group is less than the satisfaction level threshold, obtaining test sample data;

[0033] Evaluate the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values;

[0034] The agent capability assessment model and the agent grouping model are retrained based on the assessment parameter values.

[0035] In a second aspect, the present application further provides a seat grouping device, the device comprising:

[0036] A receiving module, configured to receive a seat grouping request, wherein the seat grouping request carries seats to be grouped;

[0037] A capability evaluation module, used to process the seat data corresponding to each of the seats to be grouped to obtain an indicator feature vector based on a pre-trained seat capability evaluation model;

[0038] The grouping module is used to process the indicator feature vector based on the seat grouping model obtained by training to obtain the grouping result corresponding to each seat to be grouped.

[0039] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0040] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0041] In a fifth aspect, the present application also provides a computer program product, including a computer program, which implements the steps of the above method when executed by a processor.

[0042] The above-mentioned seat grouping method, device, computer equipment, computer-readable storage medium and computer program product receive a seat grouping request, which carries seats to be grouped; based on the pre-trained seat capability evaluation model, the seat data corresponding to each seat to be grouped are processed to obtain an indicator feature vector; based on the trained seat grouping model, the indicator feature vector is processed to obtain a grouping result corresponding to each seat to be grouped, so that by combining the analysis of multi-dimensional seat-related data, the seat capability can be accurately determined, and grouping through the trained seat grouping model can reduce manual intervention, reduce operating costs, and improve the efficiency of seat resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A diagram showing an application environment of a seat grouping method in an embodiment;

[0045] Figure 2 1 is a flow chart of a method for grouping seats in one embodiment;

[0046] Figure 3 This is an overall framework diagram of a seat grouping method in one embodiment;

[0047] Figure 4 is an overall flow chart of a seat grouping method in one embodiment;

[0048] Figure 5 is a detailed flow chart of a seat grouping method in one embodiment;

[0049] Figure 6 is a structural block diagram of a seat grouping device in one embodiment;

[0050] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] In traditional technologies, customer service agent allocation usually relies on manual experience or simple matching rules, which will cause at least one of the following problems:

[0053] (1) Inefficiency: Traditional methods require a lot of manual work, which is time-consuming and error-prone, and is particularly inefficient when processing large-scale data.

[0054] (2) Uneven resource allocation: Traditional methods are highly dependent on managers’ personal experience and judgment, which may lead to over-allocation or under-allocation of resources, resulting in waste of resources or insufficient services.

[0055] (3) Lack of flexibility: Traditional rule-matching methods cannot effectively predict changes in customer demand in the future, making it difficult to make reasonable staffing plans in advance.

[0056] In order to solve the above technical problems, this application introduces a logistic regression algorithm, combines multi-dimensional features and historical data, realizes more intelligent and efficient seat management, and solves the problems of low efficiency and uneven resource allocation in traditional methods. This application can significantly improve the accuracy and efficiency of seat allocation, which is specifically manifested as follows:

[0057] (1) Improve efficiency: This application has a high degree of automation, improves the scientificity and rationality of grouping, reduces labor costs, and improves allocation efficiency.

[0058] (2) Optimize resource allocation: Improve seat resource utilization and avoid resource waste through the grouping strategy of precise algorithm models.

[0059] (3) Improved flexibility: Based on the logistic regression model, the grouping strategy can be updated regularly and dynamically adjusted according to the latest service needs and customer feedback.

[0060] The following is a detailed description of the seat grouping method provided by the present application. Specifically, the seat grouping method can be applied to the following: Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers.

[0061] The server 104 can receive the seat grouping request sent by the terminal 102, and the seat grouping request carries the seats to be grouped; based on the pre-trained seat capacity evaluation model, the seat data corresponding to each seat to be grouped is processed to obtain an indicator feature vector; based on the trained seat grouping model, the indicator feature vector is processed to obtain the grouping result corresponding to each seat to be grouped. In this way, by combining the analysis of multi-dimensional seat-related data, the seat capacity can be accurately determined, and grouping by the trained seat grouping model can reduce manual intervention, reduce operating costs, and improve the efficiency of seat resource allocation.

[0062] Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0063] In an exemplary embodiment, Figure 2 As shown, a seat grouping method is provided, which is applied to Figure 1 The server in the example is used to illustrate, including the following steps 202 to 206. Among them:

[0064] S202: receiving a seat grouping request, where the seat grouping request carries seats to be grouped.

[0065] Among them, the seat grouping request can be sent by the terminal. When the seats need to be grouped, the terminal triggers the server to perform seat grouping. The seat grouping request can be regarded as a way of manually intervening in the grouping. In this application, the seats can also be re-grouped regularly.

[0066] The seats to be grouped may be all seats or seats within a selected range. For example, seats within a corresponding position range may be selected based on the organizational structure to group seats. The range of seats to be grouped is not limited here. In other embodiments, seats to be grouped may be determined based on user selection, which may be selected one by one or in batches.

[0067] S204: Based on the pre-trained agent capability evaluation model, the agent data corresponding to each agent to be grouped is processed to obtain an indicator feature vector.

[0068] The agent capability evaluation model is a pre-trained model for evaluating agent capabilities. The agent capability evaluation model is used to extract key features from agent data and obtain a comprehensive agent score based on the weight corresponding to each key feature, that is, an indicator feature vector.

[0069] The weight corresponding to each key feature is used to characterize the importance of the key feature in the overall evaluation. The comprehensive score of the agent can be determined based on the key features in the agent data. Each key feature has a preset scoring range, and the score of each key feature needs to be within the scoring range. The comprehensive score of the agent can be obtained through the weighted average formula.

[0070] S206: Based on the trained seat grouping model, the indicator feature vector is processed to obtain a grouping result corresponding to each seat to be grouped.

[0071] Each seat corresponds to a seat comprehensive score. Based on the comprehensive scores of each seat, an indicator feature vector can be obtained. The seat grouping model is used to process each seat comprehensive score in the indicator feature vector to obtain a grouping result corresponding to each seat to be grouped.

[0072] Among them, the seat grouping model can be a logistic regression model, which is used to calculate the probability that the seat to be grouped belongs to each group. By comparing these probabilities, the group to which the grouped seat belongs can be determined. For example, the maximum value of the probabilities of various groups corresponding to a seat to be grouped can be selected, and the group corresponding to the maximum value is used as the initial group to which the seat to be grouped belongs.

[0073] The logistic regression model maps the input features of the linear combination to the probabilities of different groups through the logistic function, selects appropriate features from the preprocessed data, and uses the standardized comprehensive score of the agent to determine the group. In the actual processing, the preprocessed data is divided into a training set and a data set, the model is trained with the training set, and the model parameters are determined by the maximum likelihood estimation method, so that the model can accurately predict the group to which the agent belongs based on the input features.

[0074] For ease of understanding, combined Figure 3 As shown, Figure 3 The figure is an overall framework diagram of the agent grouping method in an embodiment. In this embodiment, the data collection module provides a standard http interface for initiating an allocation service call, querying the database table to collect agent-related historical call data, status data, service level data, etc., preliminarily sorting and classifying these original data, and storing them in the database in association with the agent number.

[0075] The data processing module is used to clean the collected data, check the data integrity and accuracy, and for missing values ​​and outliers, select the mean to fill the missing values ​​of numerical data and the mode value to fill the missing values ​​of categorical data. For outliers, statistical methods are used to identify and process them, and useless outliers are deleted. The cleaned data is standardized to make the data meet the input requirements of the logistic regression model.

[0076] The agent capability assessment model is used to extract key features from preprocessed data and assign a weight to each key feature to indicate its importance in the overall assessment. A weight value scoring range is set for each key feature, and the agent comprehensive score T can be calculated using the weighted average formula.

[0077] The principle of the agent grouping model is to map the input features of the linear combination to the probabilities of different categories through the logistic function, select appropriate features from the preprocessed data, use the standardized average problem processing time, divide the preprocessed data into training sets and data sets, train the model through historical data, and determine the model parameters through the maximum likelihood estimation method, so that the model can accurately predict the group to which the agent belongs based on the input features. Construct a logistic regression model with the group to which the agent belongs as the target variable.

[0078] The agent grouping module 603 inputs the feature vector corresponding to the agent data to be grouped into the trained logistic regression model. The model calculates the probability P (y = K|T) that the agent belongs to each group, compares these probability values, and divides the agent into the group with the highest probability. The model responds by outputting the grouping result corresponding to each agent.

[0079] Model evaluation and retraining is to query the agent evaluation table through a scheduled task, calculate the satisfaction of each skill group, set the retraining satisfaction index, and when it is lower than the satisfaction index, it is necessary to re-evaluate the model performance using the test set, adjust the model parameters according to the evaluation results, and set the retraining flag to true. The retraining flag is judged. If it is true, the model is iteratively trained and the trained model replaces the original model.

[0080] The above-mentioned seat grouping method receives a seat grouping request, which carries seats to be grouped; based on the pre-trained seat capability evaluation model, the seat data corresponding to each seat to be grouped are processed to obtain an indicator feature vector; based on the trained seat grouping model, the indicator feature vector is processed to obtain a grouping result corresponding to each seat to be grouped, so that by combining the analysis of multi-dimensional seat-related data, the seat capability can be accurately determined, and grouping through the trained seat grouping model can reduce manual intervention, reduce operating costs, and improve the efficiency of seat resource allocation.

[0081] In one of the optional embodiments, the grouping result includes the grouping probability of the seats to be grouped belonging to each group; based on the seat grouping model obtained by training, the indicator feature vector is processed to obtain the grouping result corresponding to each seat to be grouped, including: based on the grouping probability of the seats to be grouped belonging to each group, determining the initial group corresponding to the seats to be grouped; determining the grouping probability corresponding to the seats to be grouped in the same initial group; when the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial grouping, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial grouping from the initial group; when the grouping probability of the seats to be grouped meets the grouping threshold of the initial grouping, generating a grouping label corresponding to the initial group for the seats to be grouped.

[0082] The grouping results include the grouping probabilities of the seats to be grouped belonging to each group. Optionally, in the present application, the initial grouping corresponding to the seats to be grouped is determined based on the grouping results corresponding to each seat to be grouped. For example, in the present application, based on the grouping probabilities of each seat to be grouped belonging to each group, the group with the largest grouping probability can be selected as the initial group.

[0083] In this way, the seats to be grouped in each initial group and the grouping probabilities corresponding to the seats to be grouped in the initial group can be obtained.

[0084] The grouping threshold is preset and used to screen the seats to be grouped that are roughly grouped into the initial grouping. The grouping threshold may be 0.6. In other embodiments, the grouping threshold may be other values, which are only exemplified in this application.

[0085] In this application, if the probability is greater than 0.6, the seat to be grouped is determined to be able to be assigned to the initial group, and the judgment is repeated in sequence, and a group array label that meets the conditions is set for each seat to be grouped. Finally, the results of all seat allocations are summarized and calculated, and the group result set corresponding to all seats is output in response.

[0086] In one of the optional embodiments, before determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group, the step further includes: determining whether each seat to be grouped is on duty; if the seats to be grouped are not on duty, deleting the seats to be grouped; if the seats to be grouped are on duty, continuing to execute the step of determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group.

[0087] Among them, in the present application, each seat to be grouped assigned to the corresponding initial group is checked whether it is on duty. Optionally, in the present application, a corresponding number of threads can be generated based on the number of initial groups, and then each thread is processed in parallel to check whether the seat to be grouped in the corresponding initial group is on duty. If it is not on duty, the seat to be grouped is deleted from the corresponding initial group. If it is on duty, the judgment based on the group threshold is continued through these threads in parallel.

[0088] In some optional embodiments, the training method of the agent capability assessment model and the agent grouping model includes: obtaining target sample data and extracting each key feature from the target sample data; obtaining the weight assigned to each key feature; obtaining the sample indicator feature vector corresponding to each target sample data based on the weight corresponding to each key feature and the score of each key feature corresponding to each target sample data; obtaining the corresponding training grouping results based on the sample indicator feature vector, a preset logistic regression model and the total number of agent groups; training the weights assigned to each key feature and the model parameters in the preset logistic regression model based on the training grouping results and the target grouping results corresponding to the target sample data; obtaining the agent capability assessment model based on the weights assigned to each key feature obtained through training; and obtaining the agent grouping model based on the model parameters in the preset logistic regression model obtained through training.

[0089] Optionally, the target sample data in the present application is preprocessed data.

[0090] In some optional embodiments, obtaining target sample data includes: obtaining historical agent business data; cleaning the historical agent business data; performing missing data detection and abnormal data detection on the cleaned historical agent business data; processing missing data or abnormal data in the historical agent business data to obtain target sample data.

[0091] Among them, this application provides a standard http interface for initiating allocation service calls, collecting call center agent-related database tables to collect agent-related data, that is, historical agent business data. The historical agent business data includes historical call records, service time intervals, business capabilities, skill levels, service duration, customer evaluations, etc. as feature variables, and preliminarily organizes and classifies these raw data, and stores them in association with the agent numbers.

[0092] After cleaning the collected historical agent business data, check the completeness and accuracy of the historical agent business data. For missing values ​​and outliers, choose the mean to fill the missing values ​​of numerical data, and use the mode value to fill the missing values ​​of categorical data. For outliers, use statistical methods to identify and process them, choose to delete useless outliers, and standardize the cleaned data to make the data meet the input requirements of the logistic regression model, that is, to obtain the target sample data.

[0093] Subsequently, key features are extracted from the preprocessed data, i.e., the target sample data, such as the skill group service level A, business capability P, service time L, agent call answering efficiency E, average request processing time M, customer satisfaction score D, and complaint rate Q. A weight is assigned to each key feature to indicate its importance in the overall evaluation. The weights are W, A , W P , W L ,...,, the score range of the key features is defined as 1 to 10, where 10 is the highest score, and the scores are set as: S1, S2, S3,..., S n , the comprehensive score T of the agent can be calculated by the weighted average formula:

[0094]

[0095] Where i is a variable, n is the number of key features, and the characteristic vector of each comprehensive scoring index of the agent can be obtained by calculation. T = (T1, T2, ..., T m ), where m is the number of seats to be grouped.

[0096] Select appropriate features from the preprocessed data, use the standardized average problem processing time, divide the preprocessed data into 75% training set and 25% data set, and set the learning model parameter β of the logistic regression model ki , obtain the total number of skill group groups K, and use the training set data to train the model.

[0097] Use the test set to evaluate the performance of the model and train the model with historical data. Build a logistic regression model with the agent group as the target variable. The basic form of the model is:

[0098]

[0099] Where P(y=K|T) represents the probability that the seat belongs to the Kth group, T=(T1, T2, ..., T m ) is the characteristic vector of each comprehensive scoring indicator of the agent, where l is the variable of the model parameter, and its value range is greater than or equal to 1 and less than or equal to k, and i is the variable for the agent.

[0100] In the above embodiment, by combining the analysis of multi-dimensional agent-related data, the system can better predict future service needs and make reasonable personnel scheduling in advance. The grouping model strategy constructed by the logistic regression algorithm improves the utilization rate of agent resources and avoids resource waste. The system can quickly group and predict a large number of agents, which greatly saves time and labor costs compared to the traditional practice of manual evaluation and grouping one by one. Especially in scenarios with multiple skilled agents such as large customer service teams, the advantages of efficient grouping of machine learning algorithms can be highlighted. Reasonable agent grouping allows agents of different ability levels to perform their duties, achieve optimal matching of human resources and business tasks, improve overall customer service quality and efficiency, and avoid resource waste.

[0101] In one of the optional embodiments, the above method also includes: obtaining an agent evaluation form, and obtaining the satisfaction of each group based on the agent evaluation form; when the satisfaction of a group is less than a satisfaction threshold, obtaining test sample data; evaluating the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; and retraining the agent capability evaluation model and the agent grouping model based on the evaluation parameter values.

[0102] Among them, in this application, the seat evaluation table can be queried through a scheduled task to calculate the satisfaction of each skill group. When the satisfaction is lower than the satisfaction threshold, the satisfaction threshold can optionally be 98%. In other embodiments, the satisfaction threshold can also be other values. It is necessary to re-evaluate the model performance using the test set, adjust the model parameters according to the evaluation results, and set the retraining flag to true. The retraining flag is judged. If it is true, the model is iteratively trained and the trained model replaces the original model.

[0103] For ease of understanding, combined Figure 4 and Figure 5 As shown, Figure 4 FIG. 1 is an overall flow chart of a seat grouping method in an embodiment. Figure 5The detailed flow chart of the seat grouping method in one embodiment is as follows. In this embodiment, after receiving the grouping request, the seat grouping model is used to perform seat grouping, that is, the probability of the seat to be grouped being in each group is calculated, and then it is determined whether the seat to be grouped is on duty. If not, the seat data of the seat to be grouped is removed. If on duty, the seat to be grouped is judged cyclically until the initial group is determined for each seat to be grouped. After the initial group is determined for a seat to be grouped, it can be determined whether the seat to be grouped meets the grouping threshold. If not, the corresponding grouping label is not set for the seat to be grouped. If it is satisfied, the grouping label of the initial group is set for the seat to be grouped. After all the seats to be grouped are processed, the grouping result is output. It should be noted that in the present application, the seats to be grouped that are not assigned grouping labels can also be output, and the probability of each seat to be grouped belonging to each group can be output, so that the user can manually group the seats to be grouped that are not assigned grouping labels based on these probabilities.

[0104] Before processing, the data needs to be preprocessed, including cleaning, missing value checking, and outlier checking. Missing values ​​can be deleted or filled. The filling method can be seen above, and outliers can be deleted.

[0105] Then, the key features of the preprocessed data are extracted, and weights are set for each key feature, and the scoring range of each key feature is obtained, so as to set a corresponding score for each key feature. Then, the seat capacity and the above-mentioned comprehensive score are calculated by weighted average, so as to divide the training set and the test set, and obtain the total number of groups. Then, the regression model parameters are trained through the training set to construct a logistic regression model.

[0106] The training may be only for the logistic regression model, that is, to obtain the seat grouping model through training, or the seat capability evaluation model may be trained at the same time. In some optional embodiments, the seat capability evaluation model may be preset and trained separately, which is not specifically limited here.

[0107] In the above embodiment, during the agent service process, the machine learning-based system can continuously receive new data and update and optimize the model in real time. As customer needs and customer service agent status continue to change, the system can adjust the allocation strategy in a timely manner to adapt to the dynamic service environment. The system can quickly group and predict a large number of agents, which greatly saves time and labor costs compared to the traditional practice of manual evaluation and grouping one by one, especially in scenarios with multiple skilled agents such as large customer service teams, which can better highlight the advantages of efficient grouping of machine learning algorithms. Reasonable agent grouping allows agents of different ability levels to perform their duties, achieve optimal matching of human resources and business tasks, improve overall customer service quality and efficiency, and avoid waste of resources.

[0108] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0109] Based on the same inventive concept, the embodiment of the present application also provides a seat grouping device for implementing the seat grouping method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more seat grouping device embodiments provided below can refer to the limitations of the seat grouping method above, and will not be repeated here.

[0110] In an exemplary embodiment, Figure 6 As shown, a seat grouping device is provided, comprising: a receiving module 601, a capability evaluation module 602 and a grouping module 603, wherein:

[0111] A receiving module 601 is used to receive a seat grouping request, where the seat grouping request carries seats to be grouped;

[0112] A capability evaluation module 602 is used to process the seat data corresponding to each seat to be grouped based on the pre-trained seat capability evaluation model to obtain each indicator feature vector;

[0113] The grouping module 603 is used to process the indicator feature vector based on the trained seat grouping model to obtain the grouping result corresponding to each seat to be grouped.

[0114] In one of the optional embodiments, the grouping result includes the grouping probability of the seats to be grouped belonging to each group; the above-mentioned grouping module 603 is also used to determine the initial group corresponding to the seats to be grouped based on the grouping probability of the seats to be grouped belonging to each group; determine the grouping probability corresponding to the seats to be grouped in the same initial group; when the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial grouping, delete the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial grouping from the initial group; when the grouping probability of the seats to be grouped meets the grouping threshold of the initial grouping, generate a grouping label corresponding to the initial group for the seats to be grouped.

[0115] In one of the optional embodiments, the grouping module 603 is further used to determine whether each seat to be grouped is on duty; if the seat to be grouped is not on duty, the seat to be grouped is deleted; if the seat to be grouped is on duty, the step of determining the initial grouping corresponding to the seat to be grouped based on the grouping probability that the seat to be grouped belongs to each group is continued.

[0116] In one of the optional embodiments, the above-mentioned device also includes: a training module, which is used to obtain target sample data and extract each key feature from the target sample data; obtain the weight assigned to each key feature; obtain the sample indicator feature vector corresponding to each target sample data based on the weight corresponding to each key feature and the score of each key feature corresponding to each target sample data; obtain the corresponding training grouping result based on the sample indicator feature vector, the preset logistic regression model and the total number of seat groups; based on the training grouping result and the target grouping result corresponding to the target sample data, train the weights assigned to each key feature and / or the model parameters in the preset logistic regression model; obtain the seat capability evaluation model based on the weights assigned to each key feature obtained through training; and obtain the seat grouping model based on the model parameters in the preset logistic regression model obtained through training.

[0117] In one of the optional embodiments, the above-mentioned training module is also used to obtain historical agent business data; clean the historical agent business data; perform missing data detection and abnormal data detection on the cleaned historical agent business data; process business data with missing data or abnormal data in the historical agent business data to obtain target sample data.

[0118] In one of the optional embodiments, the above-mentioned training module is also used to obtain an agent evaluation form, and obtain the satisfaction of each group based on the agent evaluation form; when the satisfaction of a group is less than a satisfaction threshold, obtain test sample data; evaluate the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; and re-train the agent capability evaluation model and the agent grouping model based on the evaluation parameter values.

[0119] Each module in the above-mentioned seat grouping device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0120] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store seat grouping models and the like. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a seat grouping method is implemented.

[0121] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0122] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the following steps when executing the computer program: receiving an agent grouping request, the agent grouping request carrying agents to be grouped; based on a pre-trained agent capability evaluation model, processing the agent data corresponding to each agent to be grouped to obtain each indicator feature vector; based on the trained agent grouping model, processing the indicator feature vector to obtain a grouping result corresponding to each agent to be grouped.

[0123] In one embodiment, the grouping results involved when the processor executes the computer program include the grouping probabilities of the seats to be grouped belonging to each group; the seat grouping model based on the training implemented when the processor executes the computer program processes the indicator feature vector to obtain the grouping results corresponding to each seat to be grouped, including: determining the initial group corresponding to the seats to be grouped based on the grouping probabilities of the seats to be grouped belonging to each group; determining the grouping probabilities corresponding to the seats to be grouped in the same initial group; when the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial grouping, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial grouping from the initial group; when the grouping probability of the seats to be grouped meets the grouping threshold of the initial grouping, generating a grouping label corresponding to the initial group for the seats to be grouped.

[0124] In one embodiment, before determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group, the steps implemented by the processor when executing the computer program further include: determining whether each seat to be grouped is on duty; if the seat to be grouped is not on duty, deleting the seat to be grouped; if the seat to be grouped is on duty, continuing to execute the steps of determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group.

[0125] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining target sample data and extracting each key feature from the target sample data; obtaining the weight assigned to each key feature; obtaining the sample indicator feature vector corresponding to each target sample data based on the weight corresponding to each key feature and the score of each key feature corresponding to each target sample data; obtaining the corresponding training grouping result based on the sample indicator feature vector, a preset logistic regression model and the total number of seat groups; training the weights assigned to each key feature and / or the model parameters in the preset logistic regression model based on the training grouping result and the target grouping result corresponding to the target sample data; obtaining the agent capability evaluation model based on the weights assigned to each key feature obtained through training; and obtaining the agent grouping model based on the model parameters in the preset logistic regression model obtained through training.

[0126] In one embodiment, the acquisition of target sample data implemented when the processor executes a computer program includes: acquiring historical agent business data; cleaning the historical agent business data; performing missing data detection and abnormal data detection on the cleaned historical agent business data; processing business data with missing data or abnormal data in the historical agent business data to obtain target sample data.

[0127] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining an agent evaluation table, and obtaining the satisfaction of each group based on the agent evaluation table; when there is a group whose satisfaction is less than a satisfaction threshold, obtaining test sample data; evaluating the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; and retraining the agent capability evaluation model and the agent grouping model based on the evaluation parameter values.

[0128] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: receiving an agent grouping request, the agent grouping request carrying agents to be grouped; based on a pre-trained agent capability evaluation model, processing the agent data corresponding to each agent to be grouped to obtain each indicator feature vector; based on the trained agent grouping model, processing the indicator feature vector to obtain a grouping result corresponding to each agent to be grouped.

[0129] In one embodiment, the grouping results involved when the computer program is executed by the processor include the grouping probabilities that the seats to be grouped belong to each group; after the seat grouping model based on the training obtained is implemented when the computer program is executed by the processor, the indicator feature vector is processed to obtain the grouping results corresponding to each seat to be grouped, including: determining the initial group corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group; determining the grouping probabilities corresponding to the seats to be grouped in the same initial group; when the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial group, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial group from the initial group; when the grouping probability of the seats to be grouped meets the grouping threshold of the initial group, generating a grouping label corresponding to the initial group for the seats to be grouped.

[0130] In one embodiment, before determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group, the computer program implemented when executed by the processor also includes: determining whether each seat to be grouped is on duty; if the seat to be grouped is not on duty, deleting the seat to be grouped; if the seat to be grouped is on duty, continuing to execute the step of determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group.

[0131] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining target sample data and extracting each key feature from the target sample data; obtaining the weight assigned to each key feature; obtaining the sample indicator feature vector corresponding to each target sample data based on the weight corresponding to each key feature and the score of each key feature corresponding to each target sample data; obtaining the corresponding training grouping result based on the sample indicator feature vector, a preset logistic regression model and the total number of seat groups; training the weights assigned to each key feature and / or the model parameters in the preset logistic regression model based on the training grouping result and the target grouping result corresponding to the target sample data; obtaining the agent capability evaluation model based on the weights assigned to each key feature obtained through training; and obtaining the agent grouping model based on the model parameters in the preset logistic regression model obtained through training.

[0132] In one embodiment, the acquisition of target sample data implemented when the computer program is executed by the processor includes: acquiring historical agent business data; cleaning the historical agent business data; performing missing data detection and abnormal data detection on the cleaned historical agent business data; processing business data with missing data or abnormal data in the historical agent business data to obtain target sample data.

[0133] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining an agent evaluation table, and obtaining the satisfaction of each group based on the agent evaluation table; in the case where the satisfaction of a group is less than a satisfaction threshold, obtaining test sample data; evaluating the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; and retraining the agent capability evaluation model and the agent grouping model based on the evaluation parameter values.

[0134] In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the following steps: receiving an agent grouping request, the agent grouping request carrying agents to be grouped; based on a pre-trained agent capability evaluation model, processing the agent data corresponding to each agent to be grouped to obtain each indicator feature vector; based on the trained agent grouping model, processing the indicator feature vector to obtain a grouping result corresponding to each agent to be grouped.

[0135] In one embodiment, the grouping results involved when the computer program is executed by the processor include the grouping probabilities that the seats to be grouped belong to each group; after the seat grouping model based on the training obtained is implemented when the computer program is executed by the processor, the indicator feature vector is processed to obtain the grouping results corresponding to each seat to be grouped, including: determining the initial group corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group; determining the grouping probabilities corresponding to the seats to be grouped in the same initial group; when the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial group, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial group from the initial group; when the grouping probability of the seats to be grouped meets the grouping threshold of the initial group, generating a grouping label corresponding to the initial group for the seats to be grouped.

[0136] In one embodiment, before determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group, the computer program implemented when executed by the processor also includes: determining whether each seat to be grouped is on duty; if the seat to be grouped is not on duty, deleting the seat to be grouped; if the seat to be grouped is on duty, continuing to execute the step of determining the initial grouping corresponding to the seats to be grouped based on the grouping probability that the seats to be grouped belong to each group.

[0137] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining target sample data and extracting each key feature from the target sample data; obtaining the weight assigned to each key feature; obtaining the sample indicator feature vector corresponding to each target sample data based on the weight corresponding to each key feature and the score of each key feature corresponding to each target sample data; obtaining the corresponding training grouping result based on the sample indicator feature vector, a preset logistic regression model and the total number of seat groups; training the weights assigned to each key feature and / or the model parameters in the preset logistic regression model based on the training grouping result and the target grouping result corresponding to the target sample data; obtaining the agent capability evaluation model based on the weights assigned to each key feature obtained through training; and obtaining the agent grouping model based on the model parameters in the preset logistic regression model obtained through training.

[0138] In one embodiment, the acquisition of target sample data implemented when the computer program is executed by the processor includes: acquiring historical agent business data; cleaning the historical agent business data; performing missing data detection and abnormal data detection on the cleaned historical agent business data; processing business data with missing data or abnormal data in the historical agent business data to obtain target sample data.

[0139] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining an agent evaluation table, and obtaining the satisfaction of each group based on the agent evaluation table; in the case where the satisfaction of a group is less than a satisfaction threshold, obtaining test sample data; evaluating the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; and retraining the agent capability evaluation model and the agent grouping model based on the evaluation parameter values.

[0140] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0141] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited thereto.

[0142] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0143] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A seat grouping method, characterized in that: The method comprises: receiving a seat grouping request, wherein the seat grouping request carries seats to be grouped; Based on the pre-trained seat capability evaluation model, the seat data corresponding to each of the seats to be grouped are processed to obtain an indicator feature vector; Based on the trained seat grouping model, the indicator feature vector is processed to obtain the grouping result corresponding to each of the seats to be grouped.

2. The method according to claim 1, characterized in that The grouping result includes the grouping probability of the seats to be grouped belonging to each group; after the seat grouping model obtained through training processes the indicator feature vector to obtain the grouping result corresponding to each seat to be grouped, the method includes: Determine the initial group corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group; Determine the grouping probability corresponding to the seats to be grouped that are in the same initial group; In the case where the grouping probability of the seats to be grouped does not meet the grouping threshold of the initial grouping, deleting the seats to be grouped whose grouping probability does not meet the grouping threshold of the initial grouping from the initial grouping; When the grouping probability of the seats to be grouped meets the grouping threshold of the initial grouping, a grouping label corresponding to the initial grouping is generated for the seats to be grouped.

3. The method according to claim 2, characterized in that Before determining the initial grouping corresponding to the seats to be grouped based on the grouping probabilities that the seats to be grouped belong to each group, the method further includes: Determine whether each of the seats to be grouped is on duty; If the seat to be grouped is not on duty, deleting the seat to be grouped; In the case that the seat to be grouped is on duty, the step of determining the initial group corresponding to the seat to be grouped based on the grouping probability that the seat to be grouped belongs to each group is continued.

4. The method according to any one of claims 1 to 3, characterized in that: The training method of the agent capability evaluation model and the agent grouping model includes: Acquire target sample data, and extract key features from the target sample data; Obtaining a weight assigned to each of the key features; Based on the weights corresponding to the key features and the scores of the key features corresponding to the target sample data, a sample index feature vector corresponding to each target sample data is obtained; Based on the sample indicator feature vector, the preset logistic regression model and the total number of seat groups, a corresponding training grouping result is obtained; Based on the training grouping results and the target grouping results corresponding to the target sample data, training the weights assigned to the key features and / or the model parameters in the preset logistic regression model; Obtaining an agent capability assessment model based on the weights assigned to each of the key features obtained through training; The seat grouping model is obtained based on the model parameters in the preset logistic regression model obtained through training.

5. The method according to claim 4, characterized in that The obtaining of target sample data comprises: Obtain historical agent business data; Cleaning the historical agent business data; Performing missing data detection and abnormal data detection on the cleaned historical agent business data; The business data with missing data or abnormal data in the historical agent business data is processed to obtain target sample data.

6. The method according to claim 4, characterized in that The method further comprises: Obtaining an agent evaluation form, and obtaining satisfaction of each group based on the agent evaluation form; In the case where the satisfaction level of the group is less than the satisfaction level threshold, obtaining test sample data; Evaluate the agent capability evaluation model and the agent grouping model based on the test sample data to obtain evaluation parameter values; The agent capability evaluation model and the agent grouping model are retrained based on the evaluation parameter values.

7. A seat grouping device, characterized in that: The device comprises: A receiving module, configured to receive a seat grouping request, wherein the seat grouping request carries seats to be grouped; A capability evaluation module, used to process the seat data corresponding to each of the seats to be grouped to obtain an indicator feature vector based on a pre-trained seat capability evaluation model; The grouping module is used to process the indicator feature vector based on the seat grouping model obtained by training to obtain the grouping result corresponding to each seat to be grouped.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images