AI-based customer classification methods, devices, equipment, and storage media

Abstract

This application belongs to the fields of artificial intelligence and fintech, and relates to an AI-based customer classification method, including: acquiring customer data and constructing a topological relationship network corresponding to the customer data; inputting the topological relationship network into a graph convolutional neural network for training to obtain a trained graph convolutional neural network; acquiring a target topological relationship network corresponding to target customer data, and processing the target topological relationship network through a customer segmentation model to generate segmentation results. This application also provides an AI-based customer classification device, computer equipment, and storage medium. Furthermore, this application relates to blockchain technology, and the customer segmentation model can be stored in the blockchain. The customer classification method of this application can be applied to user identification in the financial field, requires no labeled data for training, has low time complexity, and the use of the customer segmentation model can achieve fast and accurate generation of customer segmentation results.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence technology and fintech, and in particular to artificial intelligence-based customer classification methods, devices, computer equipment, and storage media. Background Technology

[0002] Currently, fintech companies, such as insurance companies and banks, typically categorize their policyholders based on a combination of factors, including customer needs, social relationships, and perceived value. Users within the same customer group usually share certain similarities, while users in different customer groups exhibit significant differences. Based on the results of customer segmentation, fintech companies can implement different marketing strategies for different customer groups. For example, for high-value customer groups, the company can concentrate resources to develop a scientifically sound market competition strategy, thereby enhancing its competitive advantage in the market. Therefore, efficiently and accurately completing customer segmentation tasks can effectively improve the market competitiveness and market penetration of fintech companies.

[0003] Customer segmentation is widely used in finance and insurance, classifying customers into groups to meet business needs. Currently, the primary customer segmentation methods used by fintech companies involve designing deep learning-based intelligent algorithms and training the model on labeled data. However, this approach has several drawbacks: it generally requires training on labeled data, which is scarce in many companies (i.e., manually labeled user groups), and manual labeling incurs significant costs. Furthermore, this method struggles to capture the underlying clustering structures of the data during training, resulting in weak user representation vectors learned by the model and reduced accuracy in subsequent customer segmentation. Summary of the Invention

[0004] The purpose of this application is to propose a customer classification method, apparatus, computer device, and storage medium based on artificial intelligence. This addresses the problem that existing customer segmentation methods used in fintech companies generally require training with labeled data, but labeled data is scarce in these companies, and manual labeling incurs significant costs. Furthermore, these customer segmentation methods struggle to capture the underlying clustering structures of the data during training, resulting in weak user representation vectors learned by the model and reduced accuracy in subsequent customer segmentation.

[0005] To address the aforementioned technical problems, this application provides an artificial intelligence-based customer classification method, employing the following technical solution:

[0006] Acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data;

[0007] The topological relationship network is input into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed.

[0008] The first augmenting network and the second augmenting network are input into the encoder in the graph convolutional neural network, and a first low-dimensional vector representation matrix corresponding to the first augmenting network and a second low-dimensional vector representation matrix corresponding to the second augmenting network are generated in the encoder.

[0009] Construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix;

[0010] The target loss function is continuously minimized iteratively until a preset iteration stopping condition is met, at which point the iteration stops, resulting in a trained graph convolutional neural network, which is then used as a customer segmentation model.

[0011] Obtain the target topology relationship network corresponding to the target customer data to be processed, and process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers.

[0012] Furthermore, the step of inputting the topological relationship network into the construction layer of a preset graph convolutional neural network, and constructing a first augmented network and a second augmented network corresponding to the topological relationship network based on the clustering structure of the topological relationship network in the construction layer, specifically includes:

[0013] The topological relationship network is input into the construction layer of the graph convolutional neural network, and the adjacency matrix corresponding to the topological relationship network is obtained in the construction layer.

[0014] The adjacency matrix is ​​input into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network;

[0015] Determine the preset value;

[0016] Based on the clustering structure and the preset values, the edges between classes within the topological relationship network are randomly deleted to obtain the first augmented network; and...

[0017] Based on the clustering structure and the preset values, the edges between classes in the topological relationship network are randomly deleted to obtain the second augmented network.

[0018] Furthermore, the step of constructing the target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix specifically includes:

[0019] Obtain the first low-dimensional vector corresponding to the first low-dimensional vector representation matrix, and obtain the second low-dimensional vector corresponding to the second low-dimensional vector representation matrix;

[0020] A first unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector; and...

[0021] A second unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector.

[0022] The target loss function is constructed based on the first unsupervised loss function and the second unsupervised loss function.

[0023] Furthermore, the step of iteratively minimizing the target loss function until a preset iteration stopping condition is met, thereby obtaining the trained graph convolutional neural network, specifically includes:

[0024] The objective loss function is minimized iteratively, and the cumulative number of iterations is obtained.

[0025] Obtain the preset threshold number of iterations;

[0026] Determine whether the cumulative number of iterations is greater than the iteration number threshold;

[0027] If the cumulative number of iterations is greater than the iteration number threshold, the iteration stopping condition is met, the iteration stops, and the trained graph convolutional neural network is obtained.

[0028] Furthermore, after the step of determining whether the cumulative number of iterations is greater than the iteration number threshold, the method further includes:

[0029] If the cumulative number of iterations is less than the iteration number threshold, obtain the first loss function value of the current iteration process;

[0030] Obtain the second loss function value of the previous iteration process adjacent to the current iteration process;

[0031] Calculate the difference between the first loss function value and the second loss function value;

[0032] Determine whether the difference is less than a preset difference threshold;

[0033] If so, the iteration stops when the stopping condition is met, and the trained graph convolutional neural network is obtained.

[0034] Furthermore, the step of acquiring the target topology network corresponding to the target customer data to be processed, and processing the target topology network through the customer segmentation model to generate segmentation results corresponding to the target customers specifically includes:

[0035] Obtain the target topology network corresponding to the target customer data to be processed;

[0036] Input the target topology network into the customer segmentation model;

[0037] The target low-dimensional vector representation matrix corresponding to the target topological relationship network is generated using the customer segmentation model.

[0038] Based on a preset target clustering algorithm, the target low-dimensional vector representation matrix is ​​subjected to clustering partitioning to obtain the clustering partitioning result corresponding to the target low-dimensional vector representation matrix;

[0039] The clustering results are used as the segmentation results corresponding to the target customer.

[0040] Furthermore, the step of acquiring pre-collected customer data and constructing a topological relationship network corresponding to the customer data includes:

[0041] Obtain pre-collected customer data;

[0042] Extract the customer's social relationships from the customer data;

[0043] Based on the customer's social relationships, a topological relationship network corresponding to the customer is constructed.

[0044] To address the aforementioned technical problems, this application also provides an artificial intelligence-based customer classification device, which employs the following technical solution:

[0045] The first construction module is used to acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data.

[0046] A construction module is used to input the topological relationship network into a construction layer within a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed.

[0047] The generation module is used to input the first augmented network and the second augmented network into the encoder in the graph convolutional neural network, generate a first low-dimensional vector representation matrix corresponding to the first augmented network and generate a second low-dimensional vector representation matrix corresponding to the second augmented network in the encoder;

[0048] The second construction module is used to construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix;

[0049] The first processing module is used to iteratively minimize the target loss function until the preset iteration stopping condition is met, thereby obtaining a trained graph convolutional neural network, and using the trained graph convolutional neural network as a customer segmentation model.

[0050] The second processing module is used to obtain the target topology relationship network corresponding to the target customer data to be processed, and to process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers.

[0051] To address the aforementioned technical problems, this application also provides a computer device that employs the following technical solution:

[0052] Acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data;

[0053] The topological relationship network is input into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed.

[0054] The first augmenting network and the second augmenting network are input into the encoder in the graph convolutional neural network, and a first low-dimensional vector representation matrix corresponding to the first augmenting network and a second low-dimensional vector representation matrix corresponding to the second augmenting network are generated in the encoder.

[0055] Construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix;

[0056] The target loss function is continuously minimized iteratively until a preset iteration stopping condition is met, at which point the iteration stops, resulting in a trained graph convolutional neural network, which is then used as a customer segmentation model.

[0057] Obtain the target topology relationship network corresponding to the target customer data to be processed, and process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers.

[0058] To address the aforementioned technical problems, this application also provides a computer-readable storage medium, employing the technical solution described below:

[0059] Acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data;

[0060] The topological relationship network is input into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed.

[0061] The first augmenting network and the second augmenting network are input into the encoder in the graph convolutional neural network, and a first low-dimensional vector representation matrix corresponding to the first augmenting network and a second low-dimensional vector representation matrix corresponding to the second augmenting network are generated in the encoder.

[0062] Construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix;

[0063] The target loss function is continuously minimized iteratively until a preset iteration stopping condition is met, at which point the iteration stops, resulting in a trained graph convolutional neural network, which is then used as a customer segmentation model.

[0064] Obtain the target topology relationship network corresponding to the target customer data to be processed, and process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers.

[0065] Compared with the prior art, the embodiments of this application have the following main advantages:

[0066] This embodiment first acquires pre-collected customer data and constructs a topological relationship network corresponding to the customer data. Then, the topological relationship network is input into a pre-defined construction layer within a graph convolutional neural network (GCNN). In this construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed. Next, the first and second augmented networks are input into an encoder within the GCNN. The encoder generates a first low-dimensional vector representation matrix corresponding to the first augmented network and a second low-dimensional vector representation matrix corresponding to the second augmented network. Subsequently, a target loss function is constructed based on the first and second low-dimensional vector representation matrices. The target loss function is then iteratively minimized until a pre-defined iteration stopping condition is met, resulting in a trained GCNN. This trained GCNN is used as a customer segmentation model. Finally, a target topological relationship network corresponding to the target customer data is acquired, and the target topological relationship network is processed using the customer segmentation model to generate segmentation results corresponding to the target customers. This application proposes a data augmentation strategy based on clustering structure, addressing the problem that existing methods struggle to effectively utilize the inherent clustering structure of data when segmenting customers, thereby improving the accuracy of the model in user segmentation applications. The graph convolutional neural network-based method proposed in this application requires no labeled data for training and has low time complexity, making it suitable for financial applications with limited labeled data and a large customer base. This allows the trained customer segmentation model to process the target customer's data against the corresponding topological relationship network, enabling rapid and accurate generation of segmentation results corresponding to the target customer, thus improving the accuracy of the segmentation results. Attached Figure Description

[0067] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0068] Figure 1 This is an exemplary system architecture diagram to which this application can be applied;

[0069] Figure 2 A flowchart of an embodiment of the AI-based customer classification method according to this application;

[0070] Figure 3 This is a schematic diagram of a structure of an embodiment of the AI-based customer classification device according to this application;

[0071] Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0072] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0073] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0074] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0075] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0076] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0077] Terminal devices 101, 102, and 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc.

[0078] Server 105 can be a server that provides various services, such as a backend server that supports the pages displayed on terminal devices 101, 102, and 103.

[0079] It should be noted that the AI-based customer classification method provided in this application embodiment is generally executed by a server / terminal device, and correspondingly, the AI-based customer classification device is generally set in the server / terminal device.

[0080] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0081] Continue to refer to Figure 2 The flowchart illustrates an embodiment of the AI-based customer classification method according to this application. The order of steps in the flowchart can be changed, and some steps can be omitted, depending on different needs. The AI-based customer classification method provided in this application can be applied to any scenario requiring user classification, and thus can be applied to products in these scenarios, such as user classification in the financial insurance field. The AI-based customer classification method includes the following steps:

[0082] Step S201: Obtain pre-collected customer data and construct a topological relationship network corresponding to the customer data.

[0083] In this embodiment, the AI-based customer classification method operates on an electronic device (e.g., Figure 1The server / terminal device shown can acquire customer data via wired or wireless connection. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra-wideband) connections, and other currently known or future-developed wireless connection methods. The specific implementation process of acquiring pre-collected customer data and constructing a topology network corresponding to that customer data will be further described in detail in subsequent embodiments of this application, and will not be elaborated upon here.

[0084] Step S202: Input the topological relationship network into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, construct a first augmented network and a second augmented network corresponding to the topological relationship network.

[0085] In this embodiment, the specific implementation process of inputting the topological relationship network into the construction layer of a preset graph convolutional neural network, and constructing the first augmented network and the second augmented network corresponding to the topological relationship network based on the clustering structure of the topological relationship network in the construction layer, will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0086] Step S203: Input the first augmented network and the second augmented network into the encoder in the graph convolutional neural network, and generate a first low-dimensional vector representation matrix corresponding to the first augmented network and a second low-dimensional vector representation matrix corresponding to the second augmented network in the encoder.

[0087] In this embodiment, the graph convolutional neural network includes two layers. By inputting the first augmented network G1 and the second augmented network G2 into the encoder within the graph convolutional neural network, the corresponding low-dimensional vector representation matrices U and V are obtained, as shown in the following formulas: U = GCN(G1); V = GCN(G2).

[0088] Step S204: Construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix.

[0089] In this embodiment, the specific implementation process of constructing the target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0090] Step S205: Iterate continuously to minimize the target loss function until the preset iteration stopping condition is met, and then stop the iteration to obtain the trained graph convolutional neural network, and use the trained graph convolutional neural network as the customer segmentation model.

[0091] In this embodiment, the specific implementation process of iteratively minimizing the target loss function until the preset iteration stopping condition is met, thus obtaining the trained graph convolutional neural network, will be described in more detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0092] Step S206: Obtain the target topology relationship network corresponding to the target customer data to be processed, and process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers.

[0093] In this embodiment, the specific implementation process of obtaining the target topology relationship network corresponding to the target customer data to be processed, processing the target topology relationship network through the customer segmentation model, and generating segmentation results corresponding to the target customers will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0094] This application first acquires pre-collected customer data and constructs a topological relationship network corresponding to the customer data. Then, the topological relationship network is input into a pre-defined construction layer within a graph convolutional neural network (Graph Convolutional Neural Network). In this construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed. Next, the first and second augmented networks are input into an encoder within the Graph Convolutional Neural Network. The encoder generates a first low-dimensional vector representation matrix corresponding to the first augmented network and a second low-dimensional vector representation matrix corresponding to the second augmented network. Subsequently, a target loss function is constructed based on the first and second low-dimensional vector representation matrices. The target loss function is then iteratively minimized until a pre-defined iteration stopping condition is met, resulting in a trained Graph Convolutional Neural Network. This trained Graph Convolutional Neural Network is used as a customer segmentation model. Finally, a target topological relationship network corresponding to the target customer data to be processed is acquired. The target topological relationship network is then processed using the customer segmentation model to generate segmentation results corresponding to the target customers. This application proposes a data augmentation strategy based on clustering structure, addressing the problem that existing methods struggle to effectively utilize the inherent clustering structure of data when segmenting customers, thereby improving the accuracy of the model in user segmentation applications. The graph convolutional neural network-based method proposed in this application requires no labeled data for training and has low time complexity, making it suitable for applications with limited labeled data and a large customer base. This allows the trained customer segmentation model to process the target customer data's corresponding topological relationship network, enabling rapid and accurate generation of segmentation results corresponding to the target customer. This improves the accuracy of the segmentation results and provides valuable guidance for companies to efficiently, intelligently, and accurately formulate reasonable marketing strategies for different customer groups, enhancing the company's market competitiveness and market penetration.

[0095] In some alternative implementations, step S202 includes the following steps:

[0096] The topological relationship network is input into the construction layer of the graph convolutional neural network, and the adjacency matrix corresponding to the topological relationship network is obtained in the construction layer.

[0097] The adjacency matrix is ​​input into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network.

[0098] In this embodiment, the clustering algorithm is specifically the K-means clustering algorithm. By inputting the adjacency matrix into the preset clustering algorithm, the clustering structure corresponding to the topological relationship network can be obtained, as shown in the following formula: CS = {cs1 ,c s2 ,…,c sp} = Kmeans(A). Where p represents the number of clusters in the initial cluster (p > 1).

[0099] Determine the preset value.

[0100] In this embodiment, the value of the above-mentioned preset value is not specifically limited and can be set according to actual usage needs, for example, it can be set to 10%.

[0101] Based on the clustering structure and the preset values, the inter-class edges within the topological relationship network are randomly deleted to obtain the first augmented network. And,

[0102] In this embodiment, for example, if the preset value is 10%, based on the preliminary clustering structure, 10% of the inter-class connections are randomly deleted in the original topological relationship network G to obtain a new augmented network G1.

[0103] Based on the clustering structure and the preset values, the edges between classes in the topological relationship network are randomly deleted to obtain the second augmented network.

[0104] In this embodiment, the operation of deleting 10% is repeated once in the original topological relationship network G, that is, 10% of the inter-class connections are randomly deleted to obtain a new augmented network G2.

[0105] Existing methods typically employ a random augmentation strategy during the data augmentation stage, randomly deleting edges or nodes from the network. This ignores the inherent clustering structure of the topological network, resulting in weak low-dimensional customer representation vectors learned by the model and reduced accuracy in subsequent customer segmentation. This application addresses this issue by inputting the topological relationship network into the construction layer of a graph convolutional neural network. In this construction layer, an adjacency matrix corresponding to the topological relationship network is obtained. Then, the adjacency matrix is ​​input into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network. A preset value is then determined. Subsequently, based on the clustering structure and the preset value, random deletion of edges between clusters within the topological relationship network is performed to obtain the first augmented network. Finally, based on the clustering structure and the preset value, random deletion of edges between clusters within the topological relationship network is performed to obtain the second augmented network. Unlike existing technologies that randomly delete some edges in the topological relationship network, this application uses a clustering structure to selectively delete some edges in the topological relationship. As a result, the augmented network retains more topological context information, which enables the model to learn more representative low-dimensional user representation vectors. This is beneficial to improving the accuracy of customer segmentation in the subsequent generated customer segmentation model.

[0106] In some optional implementations of this embodiment, step S204 includes the following steps:

[0107] Obtain the first low-dimensional vector corresponding to the first low-dimensional vector representation matrix, and obtain the second low-dimensional vector corresponding to the second low-dimensional vector representation matrix.

[0108] In this embodiment, by inputting the first augmented network G1 and the second augmented network G2 into the encoder within the graph convolutional neural network, the corresponding low-dimensional vector representation matrices U and V are obtained, as shown in the following formulas: U = GCN(G1); V = GCN(G2). Where, u i This represents the vector in the i-th row of matrix U, and also represents the vector in the node v. i The low-dimensional vector representation learned under the condition of the first augmenting network G1. i This represents the vector in the i-th row of matrix V, and also represents the node v. i The low-dimensional vector representation learned under the condition of the second augmenting network G2. Based on this, for the same node, the two low-dimensional representations (i.e., u) in the two augmenting networks... i and v i ).

[0109] A first unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector. And,

[0110] In this embodiment, the aforementioned first unsupervised loss function specifically refers to the InfoNCE loss function, which specifically includes: Where τ is a control parameter, set to 0.2 in this scheme. ξ(u i ,v i )=sim(m(u i ),m(v i `sim(*)` calculates the cosine similarity between the two vectors within the parentheses. `m(u` i ) represents vector u i The input is the vector obtained after mapping through two layers of MLP. m(v) i Similarly.

[0111] A second unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector.

[0112] In this embodiment, the second unsupervised loss function is also the InfoNCE loss function, which can be obtained by performing corresponding data transformations based on the first unsupervised loss function, and will not be elaborated on further here.

[0113] The target loss function is constructed based on the first unsupervised loss function and the second unsupervised loss function.

[0114] In this embodiment, the target loss function can be constructed using the first unsupervised loss function and the second unsupervised loss function based on the principle of symmetry. The specific formula is as follows: Lossall = Loss(u i ,v i )+Loss(v i ,u i ). Among them, Loss(u i ,v i ) represents the first unsupervised loss function, Loss(v i ,u i Let represent the second unsupervised loss function. This application obtains a first low-dimensional vector corresponding to the first low-dimensional vector representation matrix and a second low-dimensional vector corresponding to the second low-dimensional vector representation matrix, thereby enabling the rapid and accurate construction of the target loss function based on the obtained first and second low-dimensional vectors.

[0115] This application obtains a first low-dimensional vector corresponding to the first low-dimensional vector representation matrix and a second low-dimensional vector corresponding to the second low-dimensional vector representation matrix; then constructs a first unsupervised loss function based on the first low-dimensional vector and the second low-dimensional vector; and constructs a second unsupervised loss function based on the first low-dimensional vector and the second low-dimensional vector; subsequently, it constructs the target loss function based on the first unsupervised loss function and the second unsupervised loss function.

[0116] In some alternative implementations, step S205 includes the following steps:

[0117] The target loss function is minimized iteratively, and the cumulative number of iterations is obtained.

[0118] In this embodiment, the cumulative number of iterations refers to the sum of all iterations.

[0119] Get the preset threshold number of iterations.

[0120] In this embodiment, the value of the above-mentioned iteration number threshold is not specifically limited and can be set according to actual usage requirements.

[0121] Determine whether the cumulative number of iterations is greater than the iteration number threshold.

[0122] If the cumulative number of iterations is greater than the iteration number threshold, the iteration stopping condition is met, the iteration stops, and the trained graph convolutional neural network is obtained.

[0123] In this embodiment, the model has been trained when the iteration stops.

[0124] This application iteratively minimizes the target loss function and obtains the cumulative number of iterations; then it obtains a preset iteration threshold; subsequently, it determines whether the cumulative number of iterations is greater than the iteration threshold; if the cumulative number of iterations is greater than the iteration threshold, it is determined that the iteration stopping condition is met, the iteration stops, and a trained graph convolutional neural network is obtained. Based on the comparison of the target loss function, the cumulative number of iterations, and the iteration threshold, this application can accurately train and generate the required customer segmentation model for customer segmentation.

[0125] In some optional implementations, after the step of determining whether the cumulative number of iterations is greater than the iteration number threshold, step S205 further includes the following step:

[0126] If the cumulative number of iterations is less than the iteration number threshold, obtain the first loss function value of the current iteration process.

[0127] In this embodiment, the first loss function value mentioned above refers to the loss value calculated based on the target loss function during the current iteration.

[0128] Obtain the second loss function value of the previous iteration process adjacent to the current iteration process.

[0129] In this embodiment, the second loss function value refers to the loss value calculated based on the target loss function in the previous iteration process adjacent to the current iteration process.

[0130] Calculate the difference between the first loss function value and the second loss function value.

[0131] In this embodiment, the aforementioned difference may specifically refer to the absolute value of the difference between the first loss function value and the second loss function value.

[0132] Determine whether the difference is less than a preset difference threshold.

[0133] In this embodiment, the value of the above-mentioned difference threshold is not specifically limited and can be set according to actual usage requirements.

[0134] If so, the iteration stops when the stopping condition is met, and the trained graph convolutional neural network is obtained.

[0135] In this embodiment, the model has been trained when the iteration stops.

[0136] This application, when detecting that the cumulative iteration count is less than the iteration count threshold, obtains the first loss function value of the current iteration process; then obtains the second loss function value of the previous iteration process adjacent to the current iteration process; subsequently calculates the difference between the first loss function value and the second loss function value; then determines whether the difference is less than a preset difference threshold; if so, it determines that the iteration stopping condition is met, the iteration stops, and a trained graph convolutional neural network is obtained. This application, based on the comparison of the target loss function, cumulative iteration count, iteration count threshold, and the difference between the loss function values ​​of two adjacent iteration processes, can accurately train and generate the required customer segmentation model for customer segmentation.

[0137] In some optional implementations of this embodiment, step S206 includes the following steps:

[0138] Obtain the target topology network corresponding to the target customer data to be processed.

[0139] In this embodiment, the construction process of the target topology network can refer to the construction process of the topology network described above, and will not be elaborated further here.

[0140] The target topology network is input into the customer segmentation model.

[0141] The customer segmentation model generates a target low-dimensional vector representation matrix corresponding to the target topological relationship network.

[0142] In this embodiment, a target low-dimensional vector representation matrix corresponding to the target topological relationship network can be generated by the formula H = GCN(G), where G represents the target topological relationship network, and the GCN parameters are pre-trained parameters. Therefore, this step will not update any parameters in the GCN.

[0143] The target low-dimensional vector representation matrix is ​​subjected to clustering processing based on a preset target clustering algorithm to obtain the clustering result corresponding to the target low-dimensional vector representation matrix.

[0144] In this embodiment, the aforementioned target clustering algorithm is specifically the K-means clustering algorithm. By performing clustering processing on the target low-dimensional vector representation matrix based on the K-means clustering algorithm, a clustering result corresponding to the target low-dimensional vector representation matrix can be obtained. The specific formula is as follows: C = {c1, c2, ..., c p} = Kmeans(H), where C represents the clustering result.

[0145] The clustering results are used as the segmentation results corresponding to the target customer.

[0146] This application obtains a target topological relationship network corresponding to the target customer data to be processed; then inputs the target topological relationship network into a customer segmentation model; subsequently, the customer segmentation model generates a target low-dimensional vector representation matrix corresponding to the target topological relationship network; subsequently, based on a preset target clustering algorithm, the target low-dimensional vector representation matrix is ​​clustered to obtain a clustering result corresponding to the target low-dimensional vector representation matrix, and the clustering result is used as the segmentation result corresponding to the target customer. Based on the use of the customer segmentation model, this application can quickly and accurately generate segmentation results for target customers corresponding to the target customer data to be processed, ensuring the accuracy of the generated segmentation results.

[0147] In some optional implementations of this embodiment, step S201 includes the following steps:

[0148] Obtain pre-collected customer data.

[0149] In this embodiment, for example, in the business scenario of customer classification in financial insurance, the aforementioned customer data may be pre-collected data on health insurance customers. This customer data includes at least the customer's social relationships with other customers.

[0150] Extract the customer's social relationships from the customer data.

[0151] In this embodiment, information can be extracted from the customer data based on the type of social relationship information, so as to extract the customer's social relationships from the customer data.

[0152] Based on the customer's social relationships, a topological relationship network corresponding to the customer is constructed.

[0153] In this embodiment, the aforementioned topological relationship network is a topological network G = (V, E, X) comprising the social relationships of the customers, which will be referred to as the network below. Wherein, V = {v1, v2, ..., v...} N Let} represent a network with N nodes. In this scheme, a user is represented by one node. E represents the set of all edges in the network. The existence of an edge between two nodes indicates that there is a social relationship between the two users. A represents the adjacency matrix of network G, and A ij Let v represent the element in the i-th row and j-th column of matrix A. If node v i and v j If there is an edge connecting them, then A ij =1, otherwise A ij =0. X represents the attribute matrix of the network, where each row represents the attribute vector of a node.

[0154] This application acquires pre-collected customer data; then extracts the customer's social relationships from the customer data; and subsequently constructs a topological relationship network corresponding to the customer based on the customer's social relationships, so as to quickly complete the construction of the topological relationship network, which is beneficial to quickly and accurately construct the corresponding augmented network based on the topological relationship network.

[0155] It should be emphasized that, to further ensure the privacy and security of the above customer segmentation model, the customer segmentation model can also be stored in a node of a blockchain.

[0156] The blockchain referred to in this application is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0157] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0158] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0159] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware with computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0160] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0161] Further reference Figure 3 As a response to the above Figure 2 To implement the method shown, this application provides an embodiment of an artificial intelligence-based customer classification device, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0162] like Figure 3 As shown, the AI-based customer classification device 300 described in this embodiment includes: a first construction module 301, a construction module 302, a generation module 303, a second construction module, a first processing module 305, and a second processing module 306. Wherein:

[0163] The first construction module 301 is used to acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data.

[0164] The construction module 302 is used to input the topological relationship network into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed.

[0165] The generation module 303 is used to input the first augmented network and the second augmented network into the encoder in the graph convolutional neural network, generate a first low-dimensional vector representation matrix corresponding to the first augmented network and generate a second low-dimensional vector representation matrix corresponding to the second augmented network in the encoder.

[0166] The second construction module 304 is used to construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix;

[0167] The first processing module 305 is used to iteratively minimize the target loss function until the preset iteration stopping condition is met, thereby obtaining a trained graph convolutional neural network, and using the trained graph convolutional neural network as a customer segmentation model.

[0168] The second processing module 306 is used to obtain the target topology relationship network corresponding to the target customer data to be processed, process the target topology relationship network through the customer segmentation model, and generate segmentation results corresponding to the target customers.

[0169] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0170] In some optional implementations of this embodiment, the construction module 302 includes:

[0171] The first acquisition submodule is used to input the topological relationship network into the construction layer in the graph convolutional neural network, and in the construction layer, acquire the adjacency matrix corresponding to the topological relationship network;

[0172] The first generation submodule is used to input the adjacency matrix into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network;

[0173] The first determining submodule is used to determine the preset value;

[0174] A first processing submodule is configured to randomly delete edges between classes within the topological relationship network based on the clustering structure and the preset values, thereby obtaining the first augmented network; and...

[0175] The second processing submodule is used to randomly delete the edges between classes in the topological relationship network based on the clustering structure and the preset value to obtain the second augmented network.

[0176] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0177] In some optional implementations of this embodiment, the second construction module 304 includes:

[0178] The second acquisition submodule is used to acquire the first low-dimensional vector corresponding to the first low-dimensional vector representation matrix, and to acquire the second low-dimensional vector corresponding to the second low-dimensional vector representation matrix.

[0179] A first construction submodule is used to construct a first unsupervised loss function based on the first low-dimensional vector and the second low-dimensional vector; and...

[0180] The second construction submodule is used to construct a second unsupervised loss function based on the first low-dimensional vector and the second low-dimensional vector.

[0181] The third construction submodule is used to construct the target loss function based on the first unsupervised loss function and the second unsupervised loss function.

[0182] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0183] In some optional implementations of this embodiment, the first processing module 305 includes:

[0184] The third acquisition submodule is used to continuously iterate and minimize the target loss function, and obtain the cumulative number of iterations;

[0185] The fourth submodule is used to obtain the preset iteration count threshold;

[0186] The first judgment submodule is used to determine whether the cumulative number of iterations is greater than the iteration number threshold.

[0187] The first determination submodule is used to determine that the iteration stopping condition is met if the cumulative number of iterations is greater than the iteration number threshold, and to stop the iteration to obtain the trained graph convolutional neural network.

[0188] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0189] In some optional implementations of this embodiment, the first processing module 305 further includes:

[0190] The fifth acquisition submodule is used to acquire the first loss function value of the current iteration process if the cumulative number of iterations is less than the iteration number threshold.

[0191] The sixth acquisition submodule is used to acquire the second loss function value of the previous iteration process adjacent to the current iteration process;

[0192] The calculation submodule is used to calculate the difference between the first loss function value and the second loss function value;

[0193] The second judgment submodule is used to determine whether the difference is less than a preset difference threshold.

[0194] The second decision submodule is used to determine if the iteration stopping condition is met, and then stop the iteration to obtain the trained graph convolutional neural network.

[0195] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0196] In some optional implementations of this embodiment, the second processing module 306 includes:

[0197] The seventh acquisition submodule is used to acquire the target topology relationship network corresponding to the target customer data to be processed;

[0198] The input submodule is used to input the target topological relationship network into the customer segmentation model;

[0199] The second generation submodule is used to generate a target low-dimensional vector representation matrix corresponding to the target topological relationship network through the customer segmentation model;

[0200] The partitioning submodule is used to perform clustering partitioning on the target low-dimensional vector representation matrix based on a preset target clustering algorithm, and obtain the clustering partitioning result corresponding to the target low-dimensional vector representation matrix.

[0201] The second determining submodule is used to use the clustering results as the segmentation results corresponding to the target customer.

[0202] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0203] In some optional implementations of this embodiment, the first construction module 301 includes:

[0204] The eighth submodule is used to acquire pre-collected customer data.

[0205] The extraction submodule is used to extract the customer's social relationships from the customer data.

[0206] The fourth construction submodule is used to construct a topological relationship network corresponding to the customer based on the customer's social relationships.

[0207] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based customer classification method in the aforementioned implementation method, and will not be repeated here.

[0208] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed]. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0209] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected via a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0210] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0211] The memory 41 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as the hard disk or memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Of course, the memory 41 may also include both the internal storage unit and its external storage device of the computer device 4. In this embodiment, the memory 41 is typically used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions for customer classification methods based on artificial intelligence. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or will be output.

[0212] In some embodiments, the processor 42 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions stored in the memory 41 or to process data, for example, to execute computer-readable instructions for the AI-based customer classification method.

[0213] The network interface 43 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 4 and other electronic devices.

[0214] Compared with the prior art, the embodiments of this application have the following main advantages:

[0215] This application proposes a data augmentation strategy based on clustering structure, addressing the problem that existing methods struggle to effectively utilize the inherent clustering structure of data when segmenting customers, thereby improving the accuracy of the model in user segmentation applications. The graph convolutional neural network-based method proposed in this application requires no labeled data for training and has low time complexity, making it suitable for applications with limited labeled data and a large customer base. This allows the trained customer segmentation model to process the target topological relationship network corresponding to the target customer's data, enabling rapid and accurate generation of segmentation results corresponding to the target customer, thus improving the accuracy of the segmentation results.

[0216] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based customer classification method described above.

[0217] Compared with the prior art, the embodiments of this application have the following main advantages:

[0218] This application proposes a data augmentation strategy based on clustering structure, addressing the problem that existing methods struggle to effectively utilize the inherent clustering structure of data when segmenting customers, thereby improving the accuracy of the model in user segmentation applications. The graph convolutional neural network-based method proposed in this application requires no labeled data for training and has low time complexity, making it suitable for applications with limited labeled data and a large customer base. This allows the trained customer segmentation model to process the target topological relationship network corresponding to the target customer's data, enabling rapid and accurate generation of segmentation results corresponding to the target customer, thus improving the accuracy of the segmentation results.

[0219] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0220] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A customer classification method based on artificial intelligence, characterized in that, Includes the following steps: Acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data; The topological relationship network is input into the construction layer of a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed. The first augmenting network and the second augmenting network are input into the encoder in the graph convolutional neural network, and a first low-dimensional vector representation matrix corresponding to the first augmenting network and a second low-dimensional vector representation matrix corresponding to the second augmenting network are generated in the encoder. Construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix; The target loss function is continuously minimized iteratively until a preset iteration stopping condition is met, at which point the iteration stops, resulting in a trained graph convolutional neural network, which is then used as a customer segmentation model. Obtain the target topology relationship network corresponding to the target customer data to be processed, process the target topology relationship network through the customer segmentation model, and generate segmentation results corresponding to the target customers; The step of inputting the topological relationship network into the construction layer of a preset graph convolutional neural network, and constructing a first augmented network and a second augmented network corresponding to the topological relationship network based on the clustering structure of the topological relationship network in the construction layer, specifically includes: The topological relationship network is input into the construction layer of the graph convolutional neural network, and the adjacency matrix corresponding to the topological relationship network is obtained in the construction layer. The adjacency matrix is ​​input into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network; Determine the preset value; Based on the clustering structure and the preset values, the edges between classes within the topological relationship network are randomly deleted to obtain the first augmented network; and... Based on the clustering structure and the preset values, the edges between classes in the topological relationship network are randomly deleted to obtain the second augmented network.

2. The customer classification method based on artificial intelligence according to claim 1, characterized in that, The step of constructing the target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix specifically includes: Obtain the first low-dimensional vector corresponding to the first low-dimensional vector representation matrix, and obtain the second low-dimensional vector corresponding to the second low-dimensional vector representation matrix; A first unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector; and... A second unsupervised loss function is constructed based on the first low-dimensional vector and the second low-dimensional vector. The target loss function is constructed based on the first unsupervised loss function and the second unsupervised loss function.

3. The customer classification method based on artificial intelligence according to claim 1, characterized in that, The step of iteratively minimizing the target loss function until a preset iteration stopping condition is met, thereby obtaining a trained graph convolutional neural network, specifically includes: The objective loss function is minimized iteratively, and the cumulative number of iterations is obtained. Obtain the preset threshold number of iterations; Determine whether the cumulative number of iterations is greater than the iteration number threshold; If the cumulative number of iterations is greater than the iteration number threshold, the iteration stopping condition is met, the iteration stops, and the trained graph convolutional neural network is obtained.

4. The customer classification method based on artificial intelligence according to claim 3, characterized in that, After the step of determining whether the cumulative number of iterations is greater than the iteration number threshold, the method further includes: If the cumulative number of iterations is less than the iteration number threshold, obtain the first loss function value of the current iteration process; Obtain the second loss function value of the previous iteration process adjacent to the current iteration process; Calculate the difference between the first loss function value and the second loss function value; Determine whether the difference is less than a preset difference threshold; If so, the iteration stops when the stopping condition is met, and the trained graph convolutional neural network is obtained.

5. The customer classification method based on artificial intelligence according to claim 1, characterized in that, The steps of acquiring the target topology network corresponding to the target customer data to be processed, and processing the target topology network through the customer segmentation model to generate segmentation results corresponding to the target customers, specifically include: Obtain the target topology network corresponding to the target customer data to be processed; Input the target topology network into the customer segmentation model; The target low-dimensional vector representation matrix corresponding to the target topological relationship network is generated using the customer segmentation model. Based on a preset target clustering algorithm, the target low-dimensional vector representation matrix is ​​subjected to clustering partitioning to obtain the clustering partitioning result corresponding to the target low-dimensional vector representation matrix; The clustering results are used as the segmentation results corresponding to the target customer.

6. The customer classification method based on artificial intelligence according to claim 1, characterized in that, The step of acquiring pre-collected customer data and constructing a topological relationship network corresponding to the customer data includes: Obtain pre-collected customer data; Extract the customer's social relationships from the customer data; Based on the customer's social relationships, a topological relationship network corresponding to the customer is constructed.

7. A customer classification device based on artificial intelligence, characterized in that, include: The first construction module is used to acquire pre-collected customer data and construct a topological relationship network corresponding to the customer data. A construction module is used to input the topological relationship network into a construction layer within a preset graph convolutional neural network. In the construction layer, based on the clustering structure of the topological relationship network, a first augmented network and a second augmented network corresponding to the topological relationship network are constructed. The generation module is used to input the first augmented network and the second augmented network into the encoder in the graph convolutional neural network, generate a first low-dimensional vector representation matrix corresponding to the first augmented network and generate a second low-dimensional vector representation matrix corresponding to the second augmented network in the encoder; The second construction module is used to construct a target loss function based on the first low-dimensional vector representation matrix and the second low-dimensional vector representation matrix; The first processing module is used to iteratively minimize the target loss function until the preset iteration stopping condition is met, thereby obtaining a trained graph convolutional neural network, and using the trained graph convolutional neural network as a customer segmentation model. The second processing module is used to obtain the target topology relationship network corresponding to the target customer data to be processed, and to process the target topology relationship network through the customer segmentation model to generate segmentation results corresponding to the target customers. The construction module includes: The first acquisition submodule is used to input the topological relationship network into the construction layer in the graph convolutional neural network, and in the construction layer, acquire the adjacency matrix corresponding to the topological relationship network; The first generation submodule is used to input the adjacency matrix into a preset clustering algorithm to obtain the clustering structure corresponding to the topological relationship network; The first determining submodule is used to determine the preset value; A first processing submodule is configured to randomly delete edges between classes within the topological relationship network based on the clustering structure and the preset values, thereby obtaining the first augmented network; and... The second processing submodule is used to randomly delete the edges between classes in the topological relationship network based on the clustering structure and the preset value to obtain the second augmented network.

8. A computer device comprising a memory and a processor, the memory storing computer-readable instructions, wherein the processor, when executing the computer-readable instructions, implements the steps of the artificial intelligence-based customer classification method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions that, when executed by a processor, implement the steps of the artificial intelligence-based customer classification method as described in any one of claims 1 to 6.

Images