Customer tag determination method and device, intelligent equipment and storage medium
By processing pre -processing and theme extraction of the dialogue text of customers and customer service, combined with document theme generation models and customer classification models, intelligent determining customer labels, solving the problem of low labor efficiency and poor accuracy in existing technologies, realizing more efficient And accurate customer label determination.
Patent Information
- Application Number
- CN202310936686.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-27
- Publication Date
- 2025-01-28
AI Technical Summary
In the prior art, customer tag determination relies on manual processing, and the conversation data is large and complicated, resulting in low efficiency and poor accuracy.
By obtaining the dialogue text of the target customer and customer service, enter the trained document theme generation model after pre -processing, extract the target theme, and then enter the preset customer classification model to determine the customer label.
The intelligent determination of customer labels has improved efficiency and accuracy, reduced manual intervention, and can provide customers with more accurately providing customers and controlling risks.
Smart Images
Figure CN119377721A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of user portrait technology, and in particular to a method, apparatus, intelligent device and storage medium for determining a customer tag. Background Art
[0002] In the logistics industry, in order to provide customers with high-quality services and conduct marketing risk control at the same time, customer service staff usually manually label customers to conduct subsequent customer portrait construction or customer classification. Labeling customers is also a risk control for fraud or high-risk users, which reduces malicious claims while improving the overall quality of high-quality customer service.
[0003] In the existing technology, customer labels are generally extracted from the conversation data between customer service and customers. Conversation data has the characteristic of high authenticity, and the customer labels extracted from it are of great value. However, at the same time, conversation data also has the characteristics of large data volume, redundant spoken language, diluted key information, and some untrue descriptions. If it is all processed manually, it is almost an impossible task.
[0004] How to intelligently determine customer tags and improve the efficiency and accuracy of customer tag determination is an issue that needs to be considered at present. Summary of the invention
[0005] The embodiments of the present application provide a customer tag determination method, apparatus, intelligent device, and storage medium, which can intelligently extract customer tags and improve the efficiency and accuracy of customer tag determination.
[0006] In a first aspect, an embodiment of the present application provides a method for determining a customer tag, comprising:
[0007] Get the conversation text between the target customer and the customer service;
[0008] Preprocessing the dialogue text to obtain target dialogue content;
[0009] Inputting the target conversation content into a trained document topic generation model to obtain a target topic corresponding to the target conversation content;
[0010] The target subject is input into a preset customer classification model, and based on the customer category output by the preset customer classification model, a customer label corresponding to the target customer is determined.
[0011] In a possible implementation manner of the first aspect, preprocessing the conversation text to obtain target conversation content includes:
[0012] Perform word segmentation and text vectorization on the conversation text to obtain vectorized text;
[0013] The vectorized text is cleaned and encoded to obtain the target conversation content.
[0014] In a possible implementation manner of the first aspect, before inputting the target conversation content into a trained document topic generation model to obtain a target topic corresponding to the target conversation content, the method includes:
[0015] Obtain a topic training set, where the topic training set is obtained based on historical conversation records, where the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff;
[0016] Constructing a document topic generation model, wherein model parameters of the document topic generation model have a specified parameter adjustment range;
[0017] The document topic generation model is trained by using the topic training set within the specified parameter adjustment interval.
[0018] In a possible implementation manner of the first aspect, the step of performing parameter training on the document topic generation model within the specified parameter adjustment interval using the topic training set includes:
[0019] Based on the specified parameter adjustment interval, obtaining several groups of model parameters;
[0020] Calculating the perplexity of training the document topic generation model using each set of model parameters, wherein the perplexity is determined according to the probability of the document topic generation model output during training;
[0021] The model parameters corresponding to the perplexity that meets the preset conditions are determined as the target model parameters, and the document topic generation model training is completed.
[0022] In a possible implementation manner of the first aspect, before inputting the target subject into a preset customer classification model and determining a customer label corresponding to the target customer based on a customer category output by the preset customer classification model, the method includes:
[0023] Acquire cluster samples, wherein the cluster samples include a customer classification number and a subject training set, wherein the customer classification number is determined according to a preset customer classification number interval, and the subject training set is obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff;
[0024] Obtaining a topic sample set, wherein the topic sample set is obtained by using the trained document topic generation model and the topic training set;
[0025] Taking each customer classification number as the number of cluster centers, a clustering model is constructed, and the subject sample set is clustered based on the constructed clustering model, wherein the number of the clustering models is the same as the number of customer classifications included in the customer classification number interval;
[0026] The preset customer classification model is obtained according to the clustering result of the clustering model.
[0027] In a possible implementation manner of the first aspect, obtaining the preset customer classification model according to the clustering result of the clustering model includes:
[0028] Calculate and compare the silhouette coefficient corresponding to the number of cluster centers in each clustering model;
[0029] The clustering model corresponding to the number of cluster centers with the largest silhouette coefficient is determined as the preset customer classification model.
[0030] In a possible implementation manner of the first aspect, inputting the target subject into a preset customer classification model, and determining a customer label corresponding to the target customer based on a customer category output by the preset customer classification model, includes:
[0031] Clustering the target topics using the preset customer classification model;
[0032] The customer label corresponding to the target customer is determined according to the clustering result.
[0033] In a second aspect, an embodiment of the present application provides a device for determining a customer tag, the device comprising:
[0034] A text acquisition unit, used to acquire the conversation text between the target customer and the customer service;
[0035] A text preprocessing unit, used for preprocessing the dialogue text to obtain target dialogue content;
[0036] A target topic acquisition unit, used to input the target conversation content into a trained document topic generation model to obtain a target topic corresponding to the target conversation content;
[0037] The customer label determination unit is used to input the target subject into a preset customer classification model, and determine the customer label corresponding to the target customer based on the customer category output by the preset customer classification model.
[0038] In a third aspect, an embodiment of the present application provides an intelligent device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the customer tag determination method as described in the first aspect above is implemented.
[0039] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the customer tag determination method as described in the first aspect above is implemented.
[0040] In a fifth aspect, an embodiment of the present application provides a computer program product. When the computer program product is run on a smart device, the smart device executes the customer tag determination method as described in the first aspect above.
[0041] In an embodiment of the present application, the conversation text between the target customer and the customer service is obtained, the conversation text is preprocessed to obtain the target conversation content, and then the target conversation content is input into a trained document topic generation model to obtain the target topic corresponding to the target conversation content, and then the target topic is input into a preset customer classification model, and based on the customer category output by the preset customer classification model, the customer label corresponding to the target customer is determined. There is no need to manually label the customer, and the customer label determination is intelligent, and the efficiency and accuracy of customer label determination can be effectively improved, thereby helping to provide customers with accurate services and control risks. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art 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 drawings can be obtained based on these drawings without paying creative labor.
[0043] Figure 1 is a flowchart of the implementation of the customer tag determination method provided in the embodiment of the present application;
[0044] Figure 2 This is a specific implementation flow chart of step S102 of the customer tag determination method provided in an embodiment of the present application;
[0045] Figure 3 It is a specific implementation flow chart of a training document topic generation model in the customer tag determination method provided in an embodiment of the present application;
[0046] Figure 4 It is a specific implementation flow chart of parameter adjustment and training of the document topic generation model in the customer tag determination method provided in the embodiment of the present application;
[0047] Figure 5 This is a specific implementation flow chart of obtaining a preset customer classification model in the customer tag determination method provided in the embodiment of the present application;
[0048] Figure 6 It is a specific implementation flow chart of obtaining the preset customer classification model based on the clustering result of the clustering model provided in the embodiment of the present application;
[0049] Figure 7 A specific implementation flow chart of step S104 in the customer tag determination method provided in an embodiment of the present application;
[0050] Figure 8 is a structural block diagram of a customer tag determination device provided in an embodiment of the present application;
[0051] Fig. 9 It is a schematic diagram of a smart device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0052] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.
[0053] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.
[0054] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0055] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.
[0056] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.
[0057] It should be understood that the customer tag determination method provided in the embodiment of the present application is applicable to smart devices of various types of data that need to perform user profiling and customer tag determination, which may specifically include smart devices such as mobile phones, tablet computers, wearable devices, laptops, and desktop computers. The embodiment of the present application does not impose any restrictions on the specific types of smart devices.
[0058] Figure 1 The implementation process of the customer tag determination method provided in the embodiment of the present application is shown, and the method flow includes steps S101 to S104. The specific implementation principle of each step is as follows:
[0059] Step S101: Obtain the conversation text between the target customer and the customer service.
[0060] The target customer is a customer whose customer tag is to be determined. The conversation text includes the conversation text between the target customer and the customer service in each channel.
[0061] The conversation text includes multiple conversation records between the target customer and the customer service within a period of time. The conversation records include but are not limited to voice conversation recordings, online conversation records, etc.
[0062] In some implementations, the conversation text includes date, target customer ID, customer service ID, and conversation content. The acquired conversation text is processed and cached in a specified format.
[0063] Step S102: pre-processing the dialogue text to obtain target dialogue content.
[0064] In order to improve the efficiency of conversation content processing and improve the accuracy of customer tag determination, the conversation text is preprocessed to obtain the target conversation content.
[0065] As a possible implementation of this application, Figure 2 The specific implementation process of step S101 of the customer tag determination method provided in the embodiment of the present application is shown, and is described in detail as follows:
[0066] A1: Perform word segmentation and text vectorization on the conversation text to obtain vectorized text.
[0067] In this embodiment, the dialogue text is segmented to remove stop words. In some implementations, the dialogue text is segmented based on professional terms in a specified field. For example, the dialogue text is segmented based on professional terms in the logistics field.
[0068] In one possible implementation, the TF-IDF algorithm is used for text vectorization processing. The specific definition of TF-IDF is: TF-IDF = term frequency (TF) × inverse document frequency (IDF), where inverse document frequency (IDF) = log (total number of documents in the corpus / total number of documents containing the word + 1), term frequency (TF) = the number of times a word appears in a document / the total number of words in the document.
[0069] A2: Clean and encode the vectorized text to obtain target conversation content.
[0070] The data cleaning includes data deduplication, missing value processing, outlier processing, data standardization and data conversion.
[0071] In this embodiment, one-hot encoding is used to encode the vectorized text after data cleaning to obtain the target conversation content, which helps to intelligently assist the target conversation content and lays a good foundation for subsequently determining customer tags based on the target conversation content.
[0072] Step S103: input the target conversation content into the trained document topic generation model to obtain the target topic corresponding to the target conversation content.
[0073] The document topic generation model (LDA, Latent Dirichlet Allocation) topic model is an unsupervised learning document generation model. In the embodiment of the present application, the trained document topic generation model can be used to quickly extract the target topic from the target conversation content.
[0074] As a possible implementation of this application, Figure 3 The specific implementation process of the training document topic generation model in the customer tag determination method provided in the embodiment of the present application is shown, and is described in detail as follows:
[0075] B1: Obtain a subject training set, wherein the subject training set is obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff. In this embodiment, the historical conversation texts in the subject training set are preprocessed.
[0076] B2: Construct a document topic generation model, wherein the model parameters of the document topic generation model have a specified parameter adjustment range.
[0077] The model parameters include the number of topics, a first prior distribution parameter and a second prior distribution parameter, wherein the first prior distribution parameter is specifically a document topic prior Dirichlet distribution θ d The parameter α of the second prior distribution parameter is specifically the parameter η of the prior Dirichlet distribution βk.
[0078] In this embodiment, the above-mentioned specified parameter adjustment interval can be customized by the engineer or set based on historical big data.
[0079] In some possible implementations, the above-mentioned specified parameter adjustment interval can be further limited according to a specified field.
[0080] B3: Use the topic training set to perform parameter training on the document topic generation model within the specified parameter adjustment interval.
[0081] In this embodiment, a topic training set training document topic generation model is constructed based on historical conversation texts, and model parameters are adjusted based on a specified parameter adjustment interval.
[0082] As a possible implementation of this application, Figure 4 The specific implementation process of adjusting parameters and training the document topic generation model in the customer tag determination method provided in the embodiment of the present application is shown, and is described in detail as follows:
[0083] B31: Based on the specified parameter adjustment interval, obtain several groups of model parameters. Each group of model parameters includes the number of topics n, the first prior distribution parameter and the second prior distribution parameter.
[0084] Exemplarily, the specified parameter adjustment interval of the number of topics n is [10, 15, 25, 30], the specified parameter adjustment interval of the first prior distribution parameter α is [0.001, 0.01, 0.05, 0.1, 0.2], and the specified parameter adjustment interval of the second prior distribution parameter η is [0.001, 0.01, 0.05, 0.1, 0.2]. Based on the parameter adjustment intervals of each type of model parameters, several groups of model parameters are randomly or specifiedly generated.
[0085] B32: Calculate the perplexity of training the document topic generation model using each set of model parameters, where the perplexity is determined according to the probability of the document topic generation model output during training.
[0086] In a possible implementation, the perplexity is defined as the inverse of the probability of the subject training set and normalized by the total number of samples; the specific calculation formula is as follows (1):
[0087]
[0088] Among them, PP(W) represents the perplexity, w1w2...w N represents the subject training set, P(w1w2...w N ) represents the probability of the topic training set output by the document topic generation model, and N represents the total number of training samples in the topic training set.
[0089] B33: The model parameters corresponding to the perplexity that meets the preset conditions are determined as the target model parameters, and the document topic generation model training is completed.
[0090] In this embodiment, the perplexity corresponding to each group of model parameters is sorted from high to low, where the lowest perplexity is the perplexity that meets the preset conditions, and the model parameters of the group corresponding to the lowest perplexity are determined as the target model parameters, and the document topic generation model training is completed.
[0091] In the embodiment of the present application, the perplexity is used to measure the performance of the document topic generation model during the training process. A document topic generation model with low perplexity can better predict the topic.
[0092] Step S104: input the target subject into a preset customer classification model, and determine a customer label corresponding to the target customer based on the customer category output by the preset customer classification model.
[0093] The target topic and the preset customer classification model determine the customer category, and then the customer label corresponding to the target customer is determined based on the customer category.
[0094] In some implementations, a correspondence between customer categories and customer tags is pre-established. After the customer category is determined, the customer tag can be determined based on the correspondence.
[0095] In other implementations, after the customer category of the target customer is determined, the historical logistics order information of the target customer is also obtained, and the customer tag is determined according to the customer category and the historical logistics order information.
[0096] As a possible implementation of this application, Figure 5 A specific implementation process of step S104 in the customer tag determination method provided in an embodiment of the present application is shown, and is described in detail as follows:
[0097] C1: Clustering the target topic using the preset customer classification model
[0098] C2: Determine the customer label corresponding to the target customer according to the clustering result.
[0099] In this embodiment, the clustering result includes a cluster center cluster. The center cluster is determined by clustering the target topic, and then the customer category to which the target customer belongs is determined based on the center cluster, and then the customer label is determined based on the customer category.
[0100] As a possible implementation of this application, Figure 6 A specific implementation process of obtaining a preset customer classification model in the customer tag determination method provided in an embodiment of the present application is shown, and is described in detail as follows:
[0101] D1: Obtain cluster samples, wherein the cluster samples include customer classification numbers and topic training sets, wherein the customer classification numbers are determined according to a preset customer classification number interval, and the topic training sets are obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff.
[0102] In the embodiment of the present application, the number of customer categories is used to determine the number of cluster centers, and the preset customer category number interval includes the maximum value and the minimum value of the number of cluster centers.
[0103] D2: Obtain a topic sample set, where the topic sample set is obtained using the trained document topic generation model and the topic training set.
[0104] D3: Taking each customer classification number as the number of cluster centers, construct a clustering model, and cluster the subject sample set based on the constructed clustering model, wherein the number of the clustering models is the same as the number of customer classifications included in the customer classification number interval.
[0105] In summary, in this embodiment, the clustering model may be a K-means clustering model.
[0106] D4: Obtaining the preset customer classification model according to the clustering result of the clustering model.
[0107] As a possible implementation of this application, Figure 7 A specific implementation process of obtaining the preset customer classification model according to the clustering result of the clustering model in the customer label determination method provided in the embodiment of the present application is shown, and is described in detail as follows:
[0108] D41: Calculate and compare the silhouette coefficient corresponding to the number of cluster centers in each clustering model. The silhouette coefficient is an indicator used to evaluate the quality of clustering. The calculation process of the silhouette coefficient can refer to the existing technology and will not be repeated here.
[0109] D42: Determine the clustering model corresponding to the number of cluster centers with the largest silhouette coefficient as the preset customer classification model.
[0110] In this embodiment, the maximum silhouette coefficient means the best clustering quality. Using the number of cluster centers with the largest silhouette coefficient as the preset customer classification model can effectively cluster the target topics, improve the accuracy of customer category determination, and thus improve the effectiveness and accuracy of customer label determination.
[0111] For example, taking an application scenario as an example, the K-means clustering algorithm is used to first obtain the minimum and maximum values of the number of customer classifications, so as to determine the minimum and maximum values of the number of cluster centers. For example, the minimum value is 3 and the maximum value is 10, so the K value (number of cluster centers) in K-means is in the range of [3,10]; multiple groups of historical conversation texts between customers and customer service are used as topic sample sets, and a K-means model is constructed using the topic sample set for each K value, and then the silhouette coefficient is calculated for the clustering results of each K-means model, and the K value with the largest silhouette coefficient is selected as the final number of cluster centers, and the K-means model corresponding to the K value is determined as the preset customer classification model. For example, if the optimal K is 5, it means that all customers are divided into 5 categories, and then the divided data is labeled accordingly, such as: high-quality customers, good customers, ordinary customers, suspected fraud customers, and blacklisted customers.
[0112] In this embodiment, big data and artificial intelligence technology are combined to extract topics from the conversation records between customers and customer service, and customer tags are determined by clustering, which can effectively determine high-quality customer tags, thereby defining customer groups and adopting different strategies in marketing, compensation, complaint handling, etc., and conducting risk control for high-risk customers, while also providing normal customers with better services.
[0113] In an embodiment of the present application, the conversation text between the target customer and the customer service is obtained, the conversation text is preprocessed to obtain the target conversation content, and then the target conversation content is input into a trained document topic generation model to obtain the target topic corresponding to the target conversation content, and then the target topic is input into a preset customer classification model, and based on the customer category output by the preset customer classification model, the customer label corresponding to the target customer is determined. There is no need to manually label the customer, and the customer label determination is intelligent, and the efficiency and accuracy of customer label determination can be effectively improved, thereby helping to provide customers with accurate services and control risks.
[0114] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0115] Corresponding to the customer tag determination method described in the above embodiment, Figure 8A structural block diagram of a customer tag determination device provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.
[0116] Reference Figure 8 The client tag determination device comprises: a text acquisition unit 81, a text preprocessing unit 82, a target topic acquisition unit 83, and a client tag determination unit 84, wherein:
[0117] A text acquisition unit 81 is used to acquire the conversation text between the target customer and the customer service;
[0118] A text preprocessing unit 82, used for preprocessing the dialogue text to obtain target dialogue content;
[0119] A target topic acquisition unit 83 is used to input the target conversation content into the trained document topic generation model to obtain a target topic corresponding to the target conversation content;
[0120] The customer label determination unit 84 is used to input the target subject into a preset customer classification model, and determine the customer label corresponding to the target customer based on the customer category output by the preset customer classification model.
[0121] As a possible implementation of the present application, the text preprocessing unit 82 includes:
[0122] A word segmentation and quantization module, used for performing word segmentation and text vectorization processing on the dialogue text to obtain vectorized text;
[0123] The target conversation content acquisition module is used to perform data cleaning and encoding on the vectorized text to obtain the target conversation content.
[0124] As a possible implementation manner of the present application, the above-mentioned customer tag determination device further includes:
[0125] A subject training set acquisition unit, used to acquire a subject training set, wherein the subject training set is obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff;
[0126] A model building unit, used to build a document topic generation model, wherein the model parameters of the document topic generation model have a specified parameter adjustment range;
[0127] A parameter training unit is used to use the topic training set to perform parameter training on the document topic generation model within the specified parameter adjustment interval.
[0128] As a possible implementation of the present application, the parameter adjustment training unit is specifically used for:
[0129] Based on the specified parameter adjustment interval, obtaining several groups of model parameters;
[0130] Calculating the perplexity of training the document topic generation model using each set of model parameters, wherein the perplexity is determined according to the probability of the document topic generation model output during training;
[0131] The model parameters corresponding to the perplexity that meets the preset conditions are determined as the target model parameters, and the document topic generation model training is completed.
[0132] As a possible implementation manner of the present application, the above-mentioned customer tag determination device further includes:
[0133] A cluster sample acquisition unit, used to acquire cluster samples, wherein the cluster samples include a customer classification number and a subject training set, wherein the customer classification number is determined according to a preset customer classification number interval, and the subject training set is obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff;
[0134] A topic sample acquisition unit, used to acquire a topic sample set, wherein the topic sample set is obtained by using the trained document topic generation model and the topic training set;
[0135] A clustering model construction unit, used to construct a clustering model using each customer classification number as the number of cluster centers, and cluster the subject sample set based on the constructed clustering model, wherein the number of the clustering models is the same as the number of customer classifications included in the customer classification number interval;
[0136] The customer classification model obtaining unit is used to obtain the preset customer classification model according to the clustering result of the clustering model.
[0137] As a possible implementation manner of the present application, the above-mentioned customer classification model obtaining unit is specifically used for:
[0138] Calculate and compare the silhouette coefficient corresponding to the number of cluster centers in each clustering model;
[0139] The clustering model corresponding to the number of cluster centers with the largest silhouette coefficient is determined as the preset customer classification model.
[0140] As a possible implementation of the present application, the customer tag determination unit 84 includes:
[0141] A clustering module, used to cluster the target topics using the preset customer classification model;
[0142] The label determination module is used to determine the customer label corresponding to the target customer according to the clustering result.
[0143] In an embodiment of the present application, the conversation text between the target customer and the customer service is obtained, the conversation text is preprocessed to obtain the target conversation content, and then the target conversation content is input into a trained document topic generation model to obtain the target topic corresponding to the target conversation content, and then the target topic is input into a preset customer classification model, and based on the customer category output by the preset customer classification model, the customer label corresponding to the target customer is determined. There is no need to manually label the customer, and the customer label determination is intelligent, and the efficiency and accuracy of customer label determination can be effectively improved, thereby helping to provide customers with accurate services and control risks.
[0144] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.
[0145] The present application also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, Figures 1 to 7 The steps of any customer tag determination method represented.
[0146] The embodiment of the present application also provides an intelligent device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, Figures 1 to 7 The steps of any customer tag determination method represented.
[0147] The present application also provides a computer program product, which, when executed on a server, enables the server to execute the following Figures 1 to 7 The steps of any customer tag determination method represented.
[0148] Fig. 9 Schematic diagram of a smart device provided by an embodiment of the present application. Fig. 9 As shown, the smart device 9 of this embodiment includes: a processor 90, a memory 91, and a computer program 92 stored in the memory 91 and executable on the processor 90. When the processor 90 executes the computer program 92, the steps in the above-mentioned customer tag determination method embodiments are implemented, for example Figure 2 Alternatively, when the processor 90 executes the computer program 92, the functions of each module / unit in the above-mentioned device embodiments are realized, for example Figure 8 The functions of the units 81 to 84 are shown.
[0149] Exemplarily, the computer program 92 may be divided into one or more modules / units, which are stored in the memory 91 and executed by the processor 90 to complete the present application. The one or more modules / units may be a series of computer-readable instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 92 in the smart device 9.
[0150] The smart device 9 may include, but is not limited to, a processor 90 and a memory 91. Those skilled in the art will appreciate that Fig. 9 It is only an example of the smart device 9 and does not constitute a limitation of the smart device 9. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the smart device 9 may also include input and output devices, network access devices, buses, etc.
[0151] The processor 90 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0152] The memory 91 may be an internal storage unit of the smart device 9, such as a hard disk or memory of the smart device 9. The memory 91 may also be an external storage device of the smart device 9, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the smart device 9. Further, the memory 91 may also include both an internal storage unit and an external storage device of the smart device 9. The memory 91 is used to store the computer program and other programs and data required by the smart device. The memory 91 may also be used to temporarily store data that has been output or is to be output.
[0153] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.
[0154] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.
[0155] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the device / terminal device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.
[0156] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0157] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.
Claims
1. A method for determining a customer label, characterized in that: include: Get the conversation text between the target customer and the customer service; Preprocessing the dialogue text to obtain target dialogue content; Inputting the target conversation content into a trained document topic generation model to obtain a target topic corresponding to the target conversation content; The target subject is input into a preset customer classification model, and based on the customer category output by the preset customer classification model, a customer label corresponding to the target customer is determined.
2. The method for determining a customer tag according to claim 1, characterized in that: The preprocessing of the dialogue text to obtain target dialogue content includes: Perform word segmentation and text vectorization on the conversation text to obtain vectorized text; The vectorized text is cleaned and encoded to obtain the target conversation content.
3. The method for determining a customer tag according to claim 1, characterized in that: Before inputting the target conversation content into the trained document topic generation model to obtain the target topic corresponding to the target conversation content, the method includes: Obtain a topic training set, where the topic training set is obtained based on historical conversation records, where the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff; Constructing a document topic generation model, wherein model parameters of the document topic generation model have a specified parameter adjustment range; The document topic generation model is trained by using the topic training set within the specified parameter adjustment interval.
4. The method for determining a customer tag according to claim 3, characterized in that: The using the topic training set to perform parameter training on the document topic generation model within the specified parameter adjustment interval includes: Based on the specified parameter adjustment interval, obtaining several groups of model parameters; Calculating the perplexity of training the document topic generation model using each set of model parameters, wherein the perplexity is determined according to the probability of the document topic generation model output during training; The model parameters corresponding to the perplexity that meets the preset conditions are determined as the target model parameters, and the document topic generation model training is completed.
5. The method for determining a customer tag according to claim 1, characterized in that: Before inputting the target subject into a preset customer classification model and determining the customer label corresponding to the target customer based on the customer category output by the preset customer classification model, the method includes: Acquire cluster samples, wherein the cluster samples include a customer classification number and a subject training set, wherein the customer classification number is determined according to a preset customer classification number interval, and the subject training set is obtained based on historical conversation records, wherein the historical conversation records include multiple groups of historical conversation texts between customers and customer service staff; Obtaining a topic sample set, wherein the topic sample set is obtained by using the trained document topic generation model and the topic training set; Taking each customer classification number as the number of cluster centers, a clustering model is constructed, and the subject sample set is clustered based on the constructed clustering model, wherein the number of the clustering models is the same as the number of customer classifications included in the customer classification number interval; The preset customer classification model is obtained according to the clustering result of the clustering model.
6. The method for determining a customer tag according to claim 5, characterized in that: The step of obtaining the preset customer classification model according to the clustering result of the clustering model includes: Calculate and compare the silhouette coefficient corresponding to the number of cluster centers in each clustering model; The clustering model corresponding to the number of cluster centers with the largest silhouette coefficient is determined as the preset customer classification model.
7. The method for determining a customer label according to any one of claims 1 to 6, characterized in that: The step of inputting the target subject into a preset customer classification model and determining a customer label corresponding to the target customer based on the customer category output by the preset customer classification model includes: Clustering the target topics using the preset customer classification model; The customer label corresponding to the target customer is determined according to the clustering result.
8. A customer label determination device, characterized in that: The device comprises: A text acquisition unit, used to acquire the conversation text between the target customer and the customer service; A text preprocessing unit, used for preprocessing the dialogue text to obtain target dialogue content; A target topic acquisition unit, used to input the target conversation content into a trained document topic generation model to obtain a target topic corresponding to the target conversation content; The customer label determination unit is used to input the target subject into a preset customer classification model, and determine the customer label corresponding to the target customer based on the customer category output by the preset customer classification model.
9. An intelligent device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the customer tag determination method according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the customer tag determination method according to any one of claims 1 to 7 is implemented.