Method and apparatus for identifying target account, storage medium, and electronic device

HK40077375BActive Publication Date: 2026-08-28TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
HK42022064469
Authority / Receiving Office
HK · HK
Patent Type
Patents
Current Assignee / Owner
Filing Date
2022-11-24
Publication Date
2026-08-28

AI Technical Summary

Technical Problem

In existing technologies, manual identification of malicious accounts in mobile trading platforms is inefficient and cannot effectively combat the illegal activities of malicious users.

Method used

By generating multiple interactive bipartite graphs, detecting the connection density between nodes, and determining the target account based on the detection results, the interactive bipartite graphs of heterogeneous networks are used for rapid identification.

Benefits of technology

It improves the efficiency of target account identification, enabling rapid identification of malicious accounts, reducing manual intervention time, and increasing identification efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

The application provides a target account identification method and device, a storage medium and an electronic device, and relates to the technical field of Internet. The target account identification method comprises: generating a plurality of interaction bipartite graphs according to resource interaction information between each first type account and each second type account; detecting the connection density between each node in the obtained interaction bipartite graph respectively to obtain a respective detection result; and for each candidate node, determining whether the account corresponding to the candidate node is a target account based on the total number of detection results containing the candidate node. According to the resource interaction information between each first type account and each second type account, the application generates a plurality of interaction bipartite graphs belonging to heterogeneous networks, and determines the target account in the first type account and the second type account by detecting the interaction bipartite graphs belonging to the heterogeneous networks. Through the method, the target account can be quickly identified, and the identification efficiency of the target account is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of Internet technology, and more specifically, to a method, apparatus, storage medium, and electronic device for identifying target accounts. Background Technology

[0002] With the rapid development of mobile payment, people's shopping patterns have become increasingly diversified, leading to the emergence of mobile payment methods such as scanning QR codes or facial recognition using payment apps. Through mobile transaction platforms, consumers can pay merchants, and merchants can also pay consumers.

[0003] Both merchants and consumers are users of mobile transaction platforms. The quality of users varies greatly; the vast majority are legitimate users, but some malicious users may also exist. The misconduct of malicious users, such as online fraud by merchants and consumers engaging in fraudulent transactions for merchants, can disrupt the healthy development of mobile payments. Therefore, malicious users should be punished.

[0004] Currently, accounts on the platform typically require manual identification and classification. Once an account is identified as a malicious user, it is taken into action. However, mobile trading platforms have a large number of accounts, and manual identification can only process a limited number of accounts at a time, which is also time-consuming. Therefore, the efficiency of identifying malicious accounts is low. Summary of the Invention

[0005] To address the technical problems existing in related technologies, embodiments of this application provide a target account identification method, apparatus, storage medium, and electronic device, which can improve the identification efficiency of target accounts.

[0006] To achieve the above objectives, the technical solution of this application embodiment is implemented as follows:

[0007] In a first aspect, embodiments of this application provide a target account identification method, the method comprising:

[0008] Based on the resource interaction information between each type 1 account and each type 2 account, multiple interaction bipartite graphs are generated; each interaction bipartite graph contains multiple nodes, including: type 1 nodes and type 2 nodes corresponding to type 1 accounts and type 2 accounts that have resource interaction behavior, respectively.

[0009] The connection density between each node in the obtained interactive bipartite graph is detected to obtain the detection results corresponding to each of the multiple interactive bipartite graphs; wherein the connection density between each candidate node in each detection result satisfies the set conditions.

[0010] For each obtained candidate node, perform the following operations: determine the total number of detection results containing the candidate node, and based on the total number, determine whether the account corresponding to the candidate node is the target account.

[0011] Secondly, embodiments of this application provide a target account identification device, the device comprising:

[0012] The sampling unit is used to generate multiple interaction bipartite graphs based on the resource interaction information between each first-type account and each second-type account; wherein each interaction bipartite graph contains multiple nodes, and the multiple nodes include at least: first-type nodes and second-type nodes corresponding to the first-type accounts and second-type accounts that have resource interaction behavior, respectively.

[0013] The detection unit is used to detect the connection density between each node in the obtained interactive bipartite graph and obtain the detection results corresponding to each of the multiple interactive bipartite graphs; wherein the connection density between each candidate node in each detection result satisfies the set conditions.

[0014] The identification unit is configured to perform the following operations for each obtained candidate node: determine the total number of detection results containing the candidate node, and based on the total number, determine whether the account corresponding to the candidate node is the target account.

[0015] In one optional embodiment, the sampling unit is specifically used for:

[0016] Get resource interaction information used to record resource exchange behavior between each type 1 account and each type 2 account within a set time period;

[0017] The first type of account in the resource interaction information is taken as the first type of node, and the second type of account in the resource interaction information is taken as the second type of node. Based on the resource exchange behavior between the first type of account and the second type of account, edges connecting the corresponding first type of node and the corresponding second type of node are generated to obtain the overall resource interaction graph.

[0018] Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected portion of first-type nodes and the portion of second-type nodes, as well as the edges between the selected portion of first-type nodes and the portion of second-type nodes.

[0019] Repeatedly perform the operation of selecting a portion of the first type of nodes and a portion of the second type of nodes from the overall resource interaction graph to obtain multiple interaction bipartite graphs.

[0020] In one optional embodiment, the sampling unit is specifically used for:

[0021] The resource interaction graph is transformed into a total adjacency matrix; wherein, the rows in the total adjacency matrix correspond to the second type of nodes in the resource interaction graph, and the columns in the total adjacency matrix correspond to the first type of nodes in the resource interaction graph; the matrix elements in the total adjacency matrix indicate whether there is an edge between the second type of node corresponding to its row and the first type of node corresponding to its column.

[0022] The total adjacency matrix is ​​sampled according to the set sampling rate to obtain the sub-adjacency matrix;

[0023] Based on the second type of nodes corresponding to the rows in the sub-adjacency matrix, the first type of nodes corresponding to the columns in the sub-adjacency matrix, and the matrix elements in the sub-adjacency matrix, the sub-adjacency matrix is ​​transformed into an interactive bipartite graph.

[0024] In one optional embodiment, the sampling unit is specifically used for:

[0025] Based on the interaction attribute information of the first type of account corresponding to each first type of node, each first type of node in the resource interaction master diagram is divided into the corresponding first type level; the interaction attribute information includes at least one of the following: total resource interaction amount, resource interaction frequency, or number of resource interactions.

[0026] Based on the interaction attribute information of the second type of account corresponding to each second type of node, each second type of node in the resource interaction master diagram is divided into the corresponding second type level;

[0027] According to the set sampling rate, a portion of the second-type nodes are selected from the second-type nodes contained in each second-type level, and a portion of the first-type nodes are selected from the first-type nodes contained in each first-type level, according to the same sampling rate.

[0028] In one optional embodiment, the detection unit is specifically used for:

[0029] For each interactive bipartite graph, perform the following operations:

[0030] Obtain all interactive subgraphs corresponding to the interactive bipartite graph; the number of nodes contained in each interactive subgraph is no greater than the number of nodes contained in the interactive bipartite graph.

[0031] The density of each interaction subgraph is determined based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node. The density is used to characterize the connection density between nodes in the corresponding interaction subgraph. The degree of a node is the number of edges connected to that node.

[0032] The detection result corresponding to the interaction bipartite graph is determined based on the interaction subgraph with the highest density.

[0033] In one optional embodiment, the detection unit is specifically used for:

[0034] For each interactive bipartite graph, perform the following operations:

[0035] The interactive bipartite graph is used as the bipartite graph to be processed, and the following operations are performed in a loop iteration manner until the loop termination condition is met, to obtain multiple candidate detection graphs corresponding to the interactive bipartite graph; wherein, one loop iteration process includes:

[0036] Based on the degree of each node in the bipartite graph to be processed, the number of nodes is reduced one by one to obtain multiple subgraphs to be detected; the degree of a node is the number of edges connected to the node.

[0037] The density of each subgraph to be detected is determined based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node; the density is used to characterize the connection density between nodes in the corresponding subgraph to be detected.

[0038] Nodes in the subgraph to be detected that meet the first preset condition in terms of density are added to the candidate detection graph, and the density of the candidate detection graph is determined.

[0039] Remove nodes from the subgraph to be detected whose density meets the first preset condition from the bipartite graph to be processed to obtain an updated interactive bipartite graph, and use the updated interactive bipartite graph as the bipartite graph to be processed.

[0040] Based on the density of the obtained candidate detection maps, a target detection map is determined from the multiple candidate detection maps, and based on the target detection map, the detection result corresponding to the interaction bipartite map is determined.

[0041] In one optional embodiment, the detection unit is specifically used for:

[0042] Based on the density of each candidate detection map, the second derivative of the density of each candidate detection map is determined.

[0043] Candidate detection maps whose second derivative of density meets the second preset condition are used as the target detection maps.

[0044] In one optional embodiment, the identification unit is specifically used for:

[0045] If the total number is not less than a set threshold, then the account corresponding to the candidate node is taken as the target account; or,

[0046] The detection results containing the candidate nodes are used as the target detection results;

[0047] The sum of the reciprocals of the number of nodes contained in each target detection result is taken as the detection value corresponding to the candidate node;

[0048] If the detection value corresponding to the candidate node is not less than the set detection threshold, then the account corresponding to the candidate node will be used as the target account.

[0049] Thirdly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the target account identification method of the first aspect.

[0050] Fourthly, embodiments of this application also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, the processor enables the processor to implement the target account identification method of the first aspect.

[0051] The target account identification method, apparatus, storage medium, and electronic device provided in this application generate multiple interaction bipartite graphs belonging to a heterogeneous network based on resource interaction information between each first-type account and each second-type account. By detecting the interaction bipartite graphs belonging to the heterogeneous network, the target account among the first-type and second-type accounts is determined. This method can quickly identify the target account and improve the identification efficiency. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a schematic diagram illustrating an application scenario of a target account identification method provided in an embodiment of this application.

[0054] Figure 2 A flowchart illustrating a target account identification method provided in an embodiment of this application;

[0055] Figure 3 A schematic diagram of a resource interaction overview provided in an embodiment of this application;

[0056] Figure 4 A schematic diagram illustrating a target account identification result provided in an embodiment of this application;

[0057] Figure 5 A schematic diagram illustrating another target account identification result provided in an embodiment of this application;

[0058] Figure 6 A flowchart illustrating another target account identification method provided in an embodiment of this application;

[0059] Figure 7 This is a schematic diagram illustrating the transformation of a resource interaction graph into an adjacency matrix, provided as an embodiment of this application.

[0060] Figure 8 A schematic diagram illustrating an interaction bipartite graph obtained by sampling based on an adjacency matrix, provided in an embodiment of this application;

[0061] Figure 9 for Figure 6 A flowchart illustrating step S604;

[0062] Figure 10 A structural block diagram of a target account identification device provided in an embodiment of this application;

[0063] Figure 11 A structural block diagram of an electronic device provided in an embodiment of this application;

[0064] Figure 12 This is a structural block diagram of another electronic device provided in an embodiment of this application. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0066] It should be noted that the terms "comprising" and "having" and their variations used in this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0067] The following explanations of some terms used in the embodiments of this application are provided to facilitate understanding by those skilled in the art.

[0068] (1) Homogeneous network: The concept of a network is consistent with that of a graph, which is a graph consisting of a number of given nodes and edges connecting any two nodes. A homogeneous network refers to a network in which all nodes are of the same type. For example, a network formed by user-user transactions, in which all nodes are users, is a homogeneous network.

[0069] (2) Heterogeneous network: The nodes in the network belong to at least two different types. For example, in a scholar-paper-conference network, there are three types of nodes: scholars, papers, and conferences. In the embodiments of this application, the interactive bipartite graph includes two different types of nodes: the first type of node corresponding to the first type of account and the second type of node corresponding to the second type of account. Therefore, the interactive bipartite graph belongs to a heterogeneous network.

[0070] (3) Interactive bipartite graph: The nodes in the graph belong to two different types, and there are no edges connecting nodes of the same type. For example, a network formed by merchant account-consumer account transactions, the node type in the network is either merchant account or consumer account, and there is no association between merchant accounts and consumer accounts.

[0071] The word “exemplary” as used below means “serving as an example, embodiment, or illustration.” Any embodiment illustrated as an “exemplary” need not be construed as superior to or better than other embodiments.

[0072] The terms "first" and "second" used in this document are for descriptive purposes only and should not be construed as indicating relative importance or implying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0073] The data collection, dissemination, and use in this application all comply with relevant national laws and regulations.

[0074] To better understand the technical solutions provided in the embodiments of this application, the following is a brief introduction to the application scenarios to which the technical solutions provided in the embodiments of this application are applicable. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided in the embodiments of this application can be flexibly applied according to actual needs.

[0075] This application relates to blockchain technology, a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and cryptographic algorithms. Essentially, a blockchain is a decentralized database, a chain of blocks linked together using cryptographic methods. Each block records a batch of test data used to verify the validity of the test data (anti-counterfeiting) and to generate the next block. Each block in the blockchain includes the hash value of the test data stored in that block (the hash value of this block) and the hash value of the previous block; the blocks are connected through hash values ​​to form the blockchain. Additionally, each block may also include information such as a timestamp when it was generated.

[0076] In this embodiment of the application, the transaction information of merchant accounts and consumer accounts can be saved to the blockchain in real time. The server obtains the transaction information of each merchant account and each consumer account from the blockchain and generates an interactive bipartite graph based on the transaction information.

[0077] This application also relates to Artificial Intelligence (AI) and Machine Learning (ML) technologies. Artificial intelligence utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to obtain optimal results—theories, methods, technologies, and application systems. In other words, artificial intelligence is a comprehensive technology within computer science that attempts to understand the essence of intelligence and produce new intelligent machines capable of reacting in a manner similar to human intelligence. Artificial intelligence studies the design principles and implementation methods of various intelligent machines, enabling them to possess perception, reasoning, and decision-making capabilities. Artificial intelligence technologies mainly include computer vision technology, speech processing technology, and machine learning / deep learning, among other areas.

[0078] With the research and advancement of artificial intelligence technology, AI is being studied and applied in various fields, such as smart homes, image retrieval, video surveillance, video detection, smart speakers, smart marketing, autonomous driving, drones, robots, and smart healthcare. It is believed that with the development of technology, AI will be applied in more fields and play an increasingly important role.

[0079] Machine learning is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to endow computers with intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, and inductive learning. This application's embodiments use a machine learning-based network model to generate a resource interaction master graph based on the acquired resource interaction information.

[0080] The target account identification method provided in this application can be used to identify malicious accounts on a trading platform, and can also be used to identify target accounts that meet set conditions on other platforms that include two different types of accounts. The following embodiments use the target account identification method provided in this application on a trading platform as an example for illustration.

[0081] For example, the target account identification method provided in this application embodiment can be applied to Figure 1 This is one application scenario shown. See also Figure 1 As shown, server 100 communicates with terminal devices 300 through network 200. Network 200 can be, but is not limited to, a local area network (LAN), a metropolitan area network (MAN), or a wide area network (WAN). The number of terminal devices 300 connected to server 100 can be multiple. Terminal devices 300 can transmit communication data and messages to each other through network 200 and server 100.

[0082] The terminal device 300 can be a portable device (e.g., a mobile phone, tablet computer, laptop computer, etc.), or a computer, smart screen, or personal computer (PC). The terminal device 300 can have various client applications installed, including resource interaction clients, mobile transaction clients, etc., and is an electronic device capable of displaying the various operation interfaces and objects provided by the installed client applications.

[0083] A client is software installed on a terminal device 300, such as an app installed on a mobile terminal like a smartphone, and is considered a software resource on the terminal device 300. For example, the terminal device 300 can download the client's installation package from the network, install the client using the installation package, and after installation, the client can run on the terminal device 300.

[0084] Server 100 can be a server corresponding to a trading platform. Server 100 can be a single server, a server cluster or distributed system composed of several servers, a virtualization platform, or a personal computer, a large or medium-sized computer, or a computer cluster. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. Depending on the implementation needs, the application scenarios in this application embodiment can have any number of terminal devices and servers. This application embodiment does not impose any special limitations on this.

[0085] For example, a mobile transaction client is installed on the terminal device 300. Users can register an account on the transaction platform through the mobile transaction client, such as registering a consumer account. Users can log in to the transaction platform using their registered consumer account and conduct transactions with merchant accounts registered on the transaction platform, such as making payments to merchant accounts to purchase corresponding goods or services from merchant accounts.

[0086] During each transaction between a merchant account and a consumer account, the transaction platform or server can record the transaction information. In one possible application scenario, the transaction information between each merchant account and each consumer account can be stored using cloud storage technology. Cloud storage is a new concept that extends and develops from the concept of cloud computing. A distributed cloud storage system (hereinafter referred to as a storage system) refers to a storage system that uses cluster applications, grid technology, and distributed storage file systems to aggregate a large number of storage devices of various types (storage devices are also called storage nodes) in a network to work together through application software or application interfaces to jointly provide data storage and business access functions.

[0087] In one possible application scenario, to reduce communication latency, servers 100 can be deployed in various regions, or for load balancing, different servers 100 can serve the regions corresponding to each terminal device 300. Multiple servers 100 can share data through blockchain, forming a data-sharing system. For example, terminal device 300 located at location a communicates with server 100, while terminal device 300 located at location b communicates with other servers 100.

[0088] Each server 100 in the data sharing system has a corresponding node identifier. Each server 100 can also store the node identifiers of other servers 100 in the data sharing system, so that the generated blocks can be broadcast to the other servers 100 in the data sharing system based on their node identifiers. Each server 100 can maintain a node identifier list as shown in the table below, storing the server 100 name and node identifier in the list. The node identifier can be an IP (Internet Protocol) address or any other information that can be used to identify the node. Table 1 only uses IP addresses as an example.

[0089] Table 1

[0090] Server Name Node identifier Node 1 119.115.151.174 Node 2 118.116.189.145 … … Node N 119.124.789.258

[0091] Mobile transaction platforms may harbor malicious users. Their misconduct, such as online fraud by merchants and consumers engaging in fraudulent transactions for merchants, can disrupt the healthy development of mobile payments. Currently, manually identifying and penalizing accounts as malicious is inefficient.

[0092] Based on this, embodiments of this application provide a target account identification method, apparatus, storage medium, and electronic device. In this embodiment, server 100 generates multiple interaction bipartite graphs based on resource interaction information between each first-type account and each second-type account. Each interaction bipartite graph contains multiple nodes, including at least: first-type nodes and second-type nodes corresponding to the first-type accounts and second-type accounts that have resource interaction behavior, respectively. Server 100 detects the connection density between nodes in each obtained interaction bipartite graph, obtaining detection results for each of the multiple interaction bipartite graphs; wherein the connection density between candidate nodes in each detection result satisfies a set condition. Server 100 performs the following operations for each obtained candidate node: determining the total number of detection results containing one candidate node from each candidate node; and based on the total number, determining whether the account corresponding to a candidate node is a target account.

[0093] This application embodiment generates multiple interaction bipartite graphs belonging to a heterogeneous network based on resource interaction information between each first-type account and each second-type account. By detecting these interaction bipartite graphs belonging to the heterogeneous network, the target account among the first-type and second-type accounts is determined. This method can quickly identify the target account and improve the efficiency of target account identification.

[0094] To further illustrate the technical solutions provided in the embodiments of this application, a detailed description is provided below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiments of this application provide method operation steps as shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on conventional methods or without inventive effort. For steps that do not logically have a necessary causal relationship, the execution order of these steps is not limited to the execution order provided in the embodiments of this application. In actual processing or when the device executes the method, it may be executed sequentially or in parallel according to the methods shown in the embodiments or drawings.

[0095] In one embodiment, the target account identification method provided in this application can be executed by a server or by other electronic devices with certain computing power for identifying malicious accounts. The following embodiments illustrate this method using a server executing the target account identification method as an example.

[0096] Figure 2 A flowchart of a target account identification method is shown, such as Figure 2 As shown, the method may include the following steps:

[0097] Step S201: Generate multiple interactive bipartite graphs based on the resource interaction information between each first-class account and each second-class account.

[0098] Each interaction bipartite graph contains multiple nodes, including: first-type nodes and second-type nodes corresponding to first-type accounts and second-type accounts that have resource interaction behavior, respectively. For example, the first-type account can be a merchant account, and the second-type account can be a consumer account.

[0099] In some embodiments, the server can obtain resource interaction information that records resource exchange behavior between each type of account and each type of account within a set time period. The type of accounts in the resource interaction information are designated as type 1 nodes, and the type of accounts in the resource interaction information are designated as type 2 nodes. Based on the resource exchange behavior between the type of accounts and the type of accounts, edges connecting the corresponding type 1 nodes and the corresponding type 2 nodes are generated to obtain a comprehensive resource interaction graph, such as... Figure 3 As shown. Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected first-type nodes, the selected second-type nodes, and the edges between them; repeat the operation of selecting a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph to obtain multiple interaction bipartite graphs.

[0100] For example, the transaction platform's server can obtain transaction information within one month that records the transaction behavior between various merchant accounts and various consumer accounts. The merchant accounts in the transaction information are designated as the first type of node, and the consumer accounts in the transaction information are designated as the second type of node. Based on the transaction behavior between the merchant accounts and consumer accounts, edges connecting the corresponding first type of node and the corresponding second type of node are generated to obtain the overall resource interaction graph.

[0101] In the overall resource interaction graph, any first-type node may have an edge with one or more second-type nodes. Similarly, any second-type node in the overall resource interaction graph may have an edge with one or more first-type nodes. If there is an edge between first-type node A and second-type node B in the overall resource interaction graph, it means that there was a transaction between merchant account A corresponding to first-type node A and consumer account B corresponding to second-type node B within one month.

[0102] Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected first-type nodes, the selected second-type nodes, and the edges between the selected first-type nodes and the selected second-type nodes.

[0103] Specifically, in one embodiment, the overall resource interaction graph can be transformed into an overall adjacency matrix. In this matrix, rows correspond to second-type nodes in the overall resource interaction graph, and columns correspond to first-type nodes. Matrix elements in the overall adjacency matrix indicate whether an edge exists between a second-type node in its row and a first-type node in its column.

[0104] The total adjacency matrix is ​​sampled according to the set sampling rate to obtain the sub-adjacency matrix. Based on the second type of nodes corresponding to the rows in the sub-adjacency matrix, the first type of nodes corresponding to the columns in the sub-adjacency matrix, and the matrix elements in the sub-adjacency matrix, the sub-adjacency matrix is ​​transformed into an interactive bipartite graph.

[0105] For example, the total adjacency matrix can be randomly sampled according to a set sampling rate. For instance, if the total adjacency matrix W of m×n is randomly sampled at a sampling rate of 30%, a sub-adjacency matrix of 0.3m×0.3n can be obtained.

[0106] To ensure the sampling results better reflect the distribution of merchant and consumer accounts, a hierarchical sampling method can be used. Specifically, each first-type node can be assigned to a corresponding first-type level based on the interaction attribute information of the merchant accounts corresponding to each first-type node, and each second-type node can be assigned to a corresponding second-type level based on the interaction attribute information of the second-type accounts corresponding to each second-type node. The interaction attribute information can include at least one of the following: total resource interaction amount, resource interaction frequency, or number of resource interactions.

[0107] Taking the resource interaction count as an example, this method can count the number of resource interactions, or transaction counts, for each merchant account within a set time period. Based on the resource interaction count, the first-type nodes corresponding to merchant accounts are divided into three levels: Level a, Level b, and Level c. These three levels are called the first-type levels. If the resource interaction count of merchant account A is less than the set first threshold, the first-type node corresponding to merchant account A is assigned to Level c; if the resource interaction count of merchant account A is greater than the set first threshold but less than the set second threshold, the first-type node corresponding to merchant account A is assigned to Level b; if the resource interaction count of merchant account A is greater than the set second threshold, the first-type node corresponding to merchant account A is assigned to Level a. The second threshold is greater than the first threshold.

[0108] Similarly, based on the number of resource interactions, the second-type nodes corresponding to consumer accounts are divided into three levels: level m, level n, and level l. These three levels are called the second-type levels. If the number of resource interactions for consumer account B is less than the set third threshold, the second-type node corresponding to consumer account B is assigned to level l; if the number of resource interactions for consumer account B is greater than the set third threshold but less than the set fourth threshold, the second-type node corresponding to consumer account B is assigned to level n; if the number of resource interactions for consumer account B is greater than the set fourth threshold, the second-type node corresponding to consumer account B is assigned to level m. The fourth threshold is greater than the third threshold.

[0109] After completing the hierarchical division of the first and second types of nodes, a portion of the second type nodes can be selected from each second type of node contained in each second type of node according to a set sampling rate, and a portion of the first type of nodes can be selected from each first type of node contained in each first type of node according to the same sampling rate. For example, at a sampling rate of 30%, 30% of the first type of nodes can be selected from the first type of nodes contained in level a, level b, and level c; and at a sampling rate of 30%, 30% of the second type of nodes can be selected from the second type of nodes contained in level m, level n, and level l. Determine the target columns in the total adjacency matrix corresponding to the first-class nodes selected from the three first-class levels and the target rows in the total adjacency matrix corresponding to the second-class nodes selected from the three second-class levels. Extract the matrix elements at the intersection of each target column and target row from the total adjacency matrix to obtain a sub-adjacency matrix. Based on the second-class nodes corresponding to the rows in the sub-adjacency matrix, the first-class nodes corresponding to the columns in the sub-adjacency matrix, and the matrix elements in the sub-adjacency matrix, transform the sub-adjacency matrix into an interactive bipartite graph.

[0110] In another embodiment, the first type of nodes and the second type of nodes in the overall resource interaction graph can be randomly sampled according to a set sampling rate. Based on the sampled first type of nodes and the second type of nodes, as well as the edges between each first type of node and the second type of node, an interaction bipartite graph is obtained.

[0111] Similarly, to make the sampling results more consistent with the distribution of merchant and consumer accounts, a hierarchical sampling method can be used. Specifically, based on the interaction attribute information of the first-type accounts corresponding to each first-type node, the first-type nodes in the overall resource interaction graph can be divided into corresponding first-type levels; based on the interaction attribute information of the second-type accounts corresponding to each second-type node, the second-type nodes in the overall resource interaction graph can be divided into corresponding second-type levels. The interaction attribute information includes at least one of the following: total resource interaction amount, resource interaction frequency, or number of resource interactions. Then, according to a set sampling rate, a portion of the second-type nodes are selected from the second-type nodes contained in each second-type level, and a portion of the first-type nodes are selected from the first-type nodes contained in each first-type level, according to the sampling rate. Based on the selected first-type and second-type nodes, and the edges between each first-type and second-type node, an interaction bipartite graph is generated.

[0112] Repeating any of the above sampling methods will yield multiple interactive bipartite graphs. For example, repeating the above sampling method N times will result in N interactive bipartite graphs.

[0113] Step S202: Detect the connection density between each node in the obtained interactive bipartite graph to obtain the detection results corresponding to each of the multiple interactive bipartite graphs.

[0114] In this context, the connection density between each candidate node in each detection result meets a set condition. The set condition can be that the detection result is the interaction subgraph with the highest density among all interaction subgraphs contained in the corresponding interaction bipartite graph. That is, among all interaction subgraphs contained in the corresponding interaction bipartite graph, the interaction subgraph used as the retrieval result has the densest connection density among each candidate node.

[0115] In one embodiment, the detection results corresponding to each interaction bipartite graph can be obtained as follows: For an interaction bipartite graph, all interaction subgraphs corresponding to the interaction bipartite graph are obtained. Each interaction subgraph contains first-type nodes and second-type nodes, and the number of nodes contained in the interaction subgraph is no greater than the number of nodes contained in the interaction bipartite graph. For example, suppose an interaction bipartite graph includes 6 nodes, 3 first-type nodes and 3 second-type nodes. Among them, the 3 first-type nodes are u1, u2, and u3, and the 3 second-type nodes are v1, v2, and v3. Then the interaction bipartite graph corresponds to 42 interaction subgraphs, which can be represented as {v1, u1}, {v1, u2}, {v1, u3}, {v2, u1}, {v2, u2}, {v2, u3}, {v3, u1}, {v3, u2}, {v3, u3}, {v 1. v2, u1}, {v1, v2, u2}, {v1, v2, u3}, {v1, v3, u1}, {v1, v3, u2}, {v1, v3, u3}, {v2, v3 , u1}, {v2, v3, u2}, {v2, v3, u3}, {v1, u1, u2}, {v2, u1, u2}, {v3, u1, u2}, {v1, u1, u3}, {v2, u1, u3}, {v3, u1, u3}, {v1, u2, u3}, {v2, u2, u3}, {v3, u2, u3}, {v1, v2, u1, u2}, { v1, v3, u1, u2}, {v2, v3, u1, u2}, {v1, v2, u1, u3}, {v1, v3, u1, u3}, {v2, v3, u1, u3}, {v 1. v2, u2, u3}, {v1, v3, u2, u3}, {v2, v3, u2, u3}, {v1, v2, v3, u1, u2}, {v1, v2, v3, u1, u 3}, {v1, v2, v3, u2, u3}, {v1, v2, u1, u2, u3}, {v1, v3, u1, u2, u3}, {v2, v3, u1, u2, u3}.

[0116] After obtaining all interaction subgraphs corresponding to the interaction bipartite graph, the density of each interaction subgraph can be determined based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node. Here, the degree of a node is the number of edges connected to that node; density characterizes the connection density between nodes in the corresponding interaction subgraph.

[0117] Specifically, for any interaction subgraph G, the density of the interaction subgraph G can be determined using the following density calculation formula:

[0118]

[0119] Where |u| represents the number of first-type nodes in the interaction subgraph G, |v| represents the number of second-type nodes in the interaction subgraph G, and d j Let represent the degree of the j-th first-class node in the interaction subgraph G. 'c' is a preset constant; the value of 'c' can be very small, mainly to avoid the denominator being 0. Here, the first-class node is the node corresponding to the merchant account.

[0120] After determining the density of each interaction subgraph using the above method, the interaction subgraph with the highest density is selected. Based on the interaction subgraph with the highest density, the detection result corresponding to this interaction bipartite graph is determined. Specifically, the nodes contained in the interaction subgraph with the highest density can be formed into a node set, and this node set can be used as the detection result corresponding to this interaction bipartite graph.

[0121] Perform the above operations for each interaction bipartite graph to obtain the detection results for each interaction bipartite graph.

[0122] The above example of an interaction bipartite graph with 6 nodes is just one example. In real-world applications, the number of nodes in an interaction bipartite graph may be much greater than 6. For example, each interaction bipartite graph may include 100 or 200 nodes. Therefore, the number of interaction subgraphs corresponding to each interaction bipartite graph will be quite large, and the computational workload for determining the compactness of all interaction subgraphs is very large.

[0123] To reduce computational load and conserve computing resources, in another embodiment, the detection results corresponding to each interactive bipartite graph can be obtained in the following manner, achieving the same effect as the method described above. Specifically, for an interactive bipartite graph, it can be used as the bipartite graph to be processed, and the following operations are performed iteratively until the loop termination condition is met, resulting in multiple candidate detection graphs corresponding to the interactive bipartite graph. The loop termination condition is that the number of remaining nodes in the bipartite graph to be processed is not greater than a set number. For example, when the number of remaining nodes in the bipartite graph to be processed is not greater than 2 nodes, the loop iteration process ends.

[0124] One iteration of the process may include: Based on the degree of each node in the bipartite graph to be processed, the number of nodes is gradually reduced to obtain multiple subgraphs to be tested. The degree of a node is the number of edges connected to that node. For example, assuming the bipartite graph to be processed contains 6 nodes, let's denote this bipartite graph as H6. Determine the degree of each node in H6 and find the node with the smallest degree in H6. The node with the smallest degree is removed from H6, resulting in the subgraph to be tested H5. Find the node with the smallest degree in subgraph H5 and remove it from H5, resulting in the subgraph to be tested H4. Continue in this manner, gradually reducing the number of nodes in the graph until the resulting subgraph to be tested contains only 2 nodes, resulting in the last subgraph to be tested H2. The original bipartite graph is also considered as a subgraph to be tested, resulting in a total of 5 subgraphs to be tested.

[0125] Based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node in each subgraph to be detected, the density of each subgraph to be detected is determined. Density characterizes the connection density between nodes in the corresponding subgraph to be detected. The density of each subgraph to be detected can be determined using the density calculation formula described above. Nodes in the subgraphs to be detected whose density meets a first preset condition are added to the candidate detection graph, and the density of the candidate detection graph is determined. The first preset condition can be that the density is greater than a set threshold, or that the density is the highest among all subgraphs to be detected.

[0126] Remove the nodes in the subgraph to be detected that meet the first preset density condition from the bipartite graph to be processed to obtain an updated interactive bipartite graph. Use the updated interactive bipartite graph as the bipartite graph to be processed. At this time, the number of nodes in the bipartite graph to be processed is less than the number of nodes in the original bipartite graph to be processed.

[0127] In the above iterative process, one candidate detection map is obtained with each iteration. Therefore, multiple candidate detection maps can be obtained through this iterative process. Based on the density of the obtained candidate detection maps, the target detection map is determined from the multiple candidate detection maps, and based on the target detection map, the detection result corresponding to the interaction bipartite graph is determined.

[0128] Specifically, based on the density of each candidate detection map, the second derivative of the density of each candidate detection map can be determined. The candidate detection map whose second derivative of density meets a second preset condition is then used as the target detection map. This second preset condition can be that the second derivative of density is less than a set threshold, or that the second derivative is the smallest among all candidate detection maps. After determining the target detection map, the nodes contained in the target detection map can be grouped into a node set, which is then used as the detection result corresponding to the interactive bipartite graph. The nodes in the detection result can be called candidate nodes.

[0129] Perform the above operations for each interactive bipartite graph to obtain the detection results for each interactive bipartite graph. Each detection result includes multiple candidate nodes.

[0130] Step S203: Perform the following operations for each obtained candidate node: determine the total number of detection results containing the candidate node, and based on the total number, determine whether the account corresponding to the candidate node is the target account.

[0131] In one embodiment, for each candidate node, if the total number of detection results containing that candidate node is not less than a set threshold, then the account corresponding to that candidate node is taken as the target account.

[0132] For example, suppose the threshold is set to 3, such as Figure 4 As shown, for the first type of nodes u2 and u3 corresponding to the merchant account, both nodes are candidate nodes. The total number of detection results including candidate node u2 is 2, which is less than the set threshold of 3. Therefore, the merchant account corresponding to candidate node u2 is not the target account. The total number of detection results including candidate node u3 is 3, which is not less than the set threshold of 3. Therefore, the merchant account corresponding to candidate node u3 is the target account. If a merchant account is identified as a target account, it indicates that the merchant account is a problematic merchant, such as one involved in fraud or money laundering.

[0133] like Figure 5 As shown, for the second type of nodes v1 and v5 corresponding to the consumer account, both nodes are candidate nodes. The total number of detection results, including candidate node v1, is 4, which is greater than the set threshold of 3. Therefore, the consumer account corresponding to candidate node v1 is a target account. The total number of detection results, including candidate node v5, is 3, which is equal to the set threshold of 3. Therefore, the consumer account corresponding to candidate node v5 is also a target account. If a consumer account is identified as a target account, it indicates that the consumer account is a malicious consumer, such as an account involved in money laundering or fraudulent transactions.

[0134] The principle of this method is as follows: The detection results corresponding to each interactive bipartite graph are used to detect abnormal nodes from each node. The candidate nodes with the highest connection density in the detection results indicate that each candidate node has a close connection with the nodes within the local range corresponding to the detection results. This closeness is generally much greater than the connection density between each candidate node and other nodes within the global range corresponding to the trading platform. This suggests that these candidate nodes are suspected of being abnormal and are therefore considered abnormal nodes. If a node is identified as an abnormal node by multiple detection results, it indicates a very high probability that it is an abnormal node, and the account corresponding to that node can be considered a problematic account.

[0135] In another embodiment, when determining the target account, to ensure a more balanced contribution between the detection results corresponding to the large image and the detection results corresponding to the small image, where the large image refers to the target detection image containing a large number of nodes and the small image refers to the target detection image containing a small number of nodes, for each candidate node, the detection result containing that candidate node is taken as the target detection result, and the sum of the reciprocals of the number of nodes contained in each target detection result is taken as the detection value corresponding to that candidate node. If the detection value corresponding to that candidate node is not less than a set detection threshold, then the account corresponding to that candidate node is taken as the target account.

[0136] For example, suppose there are 4 detection results for candidate node v1. The first detection result contains D nodes, the second detection result contains M nodes, the third detection result contains C nodes, and the fourth detection result contains L nodes. Then the reciprocals of the number of nodes in the four detection results are: The detection value corresponding to candidate node v1 is the sum of the reciprocals: If the detected value is greater than or equal to the set detection threshold, then the account corresponding to the candidate node v1 is taken as the target account.

[0137] The target account identification method provided in this application generates multiple interaction bipartite graphs belonging to heterogeneous networks based on resource interaction information between each first-type account and each second-type account. By detecting the interaction bipartite graphs belonging to heterogeneous networks, malicious accounts in the first-type and second-type accounts are identified. Compared with related technologies that analyze homogeneous networks based on spectral methods to identify malicious accounts, the algorithm is relatively simple, less time-consuming, can identify malicious accounts in real time in online environments, has high accuracy, and effectively combats identified malicious users.

[0138] To better understand the target account identification method in the embodiments of this application, Figure 6 This illustrates a specific implementation process of an embodiment of this application. For example... Figure 6 As shown, the process includes the following steps:

[0139] Step S601: Obtain transaction information of transactions conducted through a designated trading platform within a set time period, and generate a resource interaction master map based on the obtained transaction information.

[0140] The designated trading platform can be any internet trading platform, through which consumer and merchant accounts can conduct transactions. For example, the designated trading platform could be a shopping platform. After merchants join the shopping platform, they can sell their goods through it, and consumers can purchase these goods and pay the merchants through the platform. During these transactions, the shopping platform can record transaction information, including consumer account information, merchant account information, and transaction amount.

[0141] This application embodiment can acquire transaction information from multiple transaction processes recorded on a specified transaction platform within a set time period. This transaction information includes multiple merchant accounts and multiple consumer accounts, as well as payment records between different merchant accounts and different consumer accounts. Payment records can include two types of payment records: commercial payments and enterprise payments. Commercial payments and enterprise payments form reverse payments; commercial payments refer to payments made by a consumer account to a merchant account, and enterprise payments refer to payments made by a merchant account to a consumer account.

[0142] Based on the transaction information from the aforementioned multiple transaction processes, a resource interaction master graph can be generated to represent the transaction relationship between consumer accounts and merchant accounts. This resource interaction master graph is a bipartite graph, where nodes represent consumer accounts and merchant accounts. For example, Figure 3 This diagram illustrates a resource interaction master map, which can also be referred to as a merchant-consumer transaction network. For example... Figure 3 As shown, nodes v1 to v5 represent merchant accounts and can be called merchant nodes, while nodes u1 to u5 represent consumer accounts and can be called consumer nodes. The connection between a merchant node and a consumer node can be called an edge, indicating that there is a transaction relationship between the corresponding merchant account and the corresponding consumer account. For example, the connection between merchant node v1 and consumer node u1 indicates that merchant account v1 and consumer account u1 have transacted within a set time period, and a transaction relationship exists between them; there is no connection between merchant node v4 and consumer node u3, indicating that merchant account v4 and consumer account u3 have not transacted within the set time period, and no transaction relationship exists between them.

[0143] Figure 3The number of merchant nodes and consumer nodes shown is equal. In other embodiments, the number of merchant nodes and consumer nodes may not be equal. For example, the overall resource interaction graph includes 100 merchant nodes and 150 consumer nodes.

[0144] Step S602: Convert the resource interaction graph into an adjacency matrix.

[0145] An adjacency matrix can be a two-dimensional array, where rows represent merchant accounts and columns represent consumer accounts. Elements in the adjacency matrix are either 0 or 1. "0" indicates that there is no transaction relationship between the merchant account in that row and the consumer account in that column, while "1" indicates that there is a transaction relationship between the two accounts.

[0146] For example, suppose the adjacency matrix W obtained from the resource interaction graph is an m×n matrix, where the number of rows m indicates the existence of m merchant accounts, corresponding to merchant account v1 to merchant account v2 respectively. m The number of columns n indicates that there are n consumer accounts, corresponding to merchant accounts u1 to u2 respectively. n If the element in the first row and first column of the adjacency matrix W is 1, it indicates that there is a transaction relationship between merchant account v1 and consumer account u1, and there is a connection between merchant node v1 and consumer node u1 in the overall resource interaction graph. For example, Figure 3 The resource interaction master diagram shown can be transformed into Figure 7 The adjacency matrix shown.

[0147] Step S603: Randomly sample the adjacency matrix multiple times according to the set sampling rate to obtain multiple sub-adjacency matrices, and determine the interactive bipartite graph corresponding to each sub-adjacency matrix.

[0148] In one embodiment, assuming a sampling rate of 30%, the adjacency matrix W is randomly sampled at a sampling rate of 30% to obtain a 0.3m × 0.3n sub-adjacency matrix. Specifically, 0.3m rows are randomly selected from the m rows of the adjacency matrix W, or 0.3m merchant accounts are randomly selected from the m merchant accounts corresponding to the adjacency matrix W; 0.3n columns are randomly selected from the n columns of the adjacency matrix W, or 0.3n consumer accounts are randomly selected from the n consumer accounts corresponding to the adjacency matrix W. For example, assuming the adjacency matrix W is a 200 × 100 matrix, 60 rows are randomly selected from the 200 rows of the adjacency matrix W, and 30 columns are randomly selected from the 100 columns of the adjacency matrix W. The matrix elements at the intersection of the selected 60 rows and the selected 30 columns are extracted from the adjacency matrix W to form the sub-adjacency matrix.

[0149] Repeat the operation of randomly sampling the adjacency matrix W at a sampling rate of 30% N times to obtain N sub-adjacency matrices, which are respectively Based on the N sub-adjacency matrices, determine the corresponding interactive bipartite graph for each sub-adjacency matrix. Each interactive bipartite graph includes 0.3m merchant nodes corresponding to merchant accounts and 0.3n consumer nodes corresponding to consumer accounts. Whether there is a connection between a merchant node and a consumer node can be determined by the matrix element at the intersection of the row corresponding to the merchant account and the column corresponding to the consumer account in the corresponding sub-adjacency matrix.

[0150] In another embodiment, assume the sampling rate is set to 60%. Figure 8 As shown, a 5×5 adjacency matrix is ​​randomly sampled at a sampling rate of 60%, resulting in a 3×3 sub-adjacency matrix. This sub-adjacency matrix contains rows corresponding to merchant nodes v1, v2, and v3 and columns corresponding to consumer nodes u1, u2, and u4. Based on this sub-adjacency matrix, the determined interaction bipartite graph is as follows. Figure 8 As shown, in this interactive bipartite graph, the degree of consumer node u4 is 0.

[0151] Step S604: Based on the connection relationships between nodes in each interactive bipartite graph, determine the detection results corresponding to each interactive bipartite graph.

[0152] Specifically, assume that N interactive bipartite graphs are obtained in step S603. For each interactive bipartite graph, respectively, according to... Figure 9 The method shown involves the following steps:

[0153] Step S6041: Use the interactive bipartite graph as the initial bipartite graph to be processed, and set the dataset S respectively. d Dataset v d and dataset u d , of which S d =(u d ∪v d Let dataset v d and dataset u d It is an empty set.

[0154] Step S6042: Based on the degree of each node in the bipartite graph to be processed, the number of nodes in the bipartite graph to be processed is reduced one by one to obtain multiple subgraphs to be detected.

[0155] The bipartite graph to be processed in this step can be the initial bipartite graph in step S6041, or the updated bipartite graph in step S6043 below. If the number of nodes in the bipartite graph to be processed is n, this number is the total number of merchant nodes and consumer nodes in the bipartite graph to be processed, that is, the bipartite graph to be processed includes n nodes in total. Let this bipartite graph to be processed be denoted as H.n Determine H n The degree of each node in H, and from H n Find the node with the smallest degree. The degree of a node is the number of edges connected to that node. Remove the node with the smallest degree from H. n Delete it to obtain the subgraph H to be detected. n-1 Find the subgraph H to be detected n-1 Find the node with the smallest degree in the middle and remove it from the subgraph H to be detected. n-1 Delete it to obtain the subgraph H to be detected. n-2 This process continues, gradually reducing the number of nodes in the graph until the resulting subgraph to be detected contains only two nodes, thus obtaining the final subgraph to be detected, H2. Assume the subgraph to be detected is H... i If the node with the smallest degree is m, then the subgraph to be detected, H, is obtained after deleting this node. i-1 It can be represented as H i-1 =H i -{m}.

[0156] The bipartite graph H to be processed n Also considered as a subgraph to be detected, the above process can yield a total of n-1 subgraphs H to be detected. i , i = n, ... 2.

[0157] Step S6043: Determine the density of each subgraph to be detected, select the subgraph with the highest density as the target detection subgraph, and add the merchant nodes contained in the target detection subgraph to the dataset v. d Add the consumer nodes contained in the object detection subgraph to the dataset u d Then, the nodes and edges contained in the object detection subgraph are removed from the bipartite graph to be processed, resulting in an updated bipartite graph to be processed.

[0158] For any bipartite graph G, the compactness of G can be determined by the following formula:

[0159]

[0160] Where |u| represents the number of consumer nodes in the bipartite graph G, |v| represents the number of merchant nodes in the bipartite graph G, and d j Let c represent the degree of the j-th merchant node in the bipartite graph G. c is a preset constant, and the value of c can be very small, mainly to avoid the denominator being 0.

[0161] Based on the above formula, the density of each sub-image to be detected can be determined, denoted as φ(H2), φ(H3), ..., φ(H2). n Determine φ(H2)~φ(H). nThe maximum value in G(H) is used to select the sub-image with the highest density as the target detection sub-image, denoted as G(H). i )=(u i ∪v i ,ε i ), where ε i G(H) represents the object detection subgraph. i The edges in the bipartite graph to be processed that are connected to the target detection subgraph G(H) are the edges in the bipartite graph to be processed. i The edges connecting each node contained in the array. Let u d =u d ∪u i ,v d =v d ∪v i The updated dataset S is obtained. d Remove the object detection sub-map G(H) from the bipartite image to be processed. i The nodes and edges contained in the graph are used to obtain the updated bipartite graph to be processed.

[0162] Step S6044: Determine the updated dataset S based on the object detection subgraph. d The corresponding bipartite graph G(S) d ), and the bipartite graph G(S) d Using G(S) as a candidate detection map, the density φ(G(S) of the candidate detection map is calculated. d The second derivative of )) is denoted as Δ 2 φ(G(S d )).

[0163] Step S6045: Determine whether the number of nodes in the updated bipartite graph to be processed is less than the set number; if yes, proceed to step S6046; if no, return to step S6042.

[0164] Step S6046: Select the dataset S with the smallest second derivative of compactness. d As the detection result of the interactive bipartite graph.

[0165] For the updated bipartite graph to be processed, repeat steps S6042 to S6044 until the updated bipartite graph contains fewer than two nodes. During the repeated execution of these steps, the dataset S... d The dataset S grows continuously, and with each growth, an updated dataset S is determined. d The second derivative of the density of the corresponding candidate detection maps is used to find the dataset S with the smallest second derivative of the density. dAs the detection result of this interactive bipartite graph, this detection result can also be called the detection result corresponding to the interactive bipartite graph, and the nodes in the detection result are called candidate nodes. Similar to the minimum slope at the vertex of the curve corresponding to a quadratic equation, the minimum second derivative of the density indicates that the corresponding dataset S... d The density of the dataset S reaches its peak, indicating that the dataset S... d The goal is to maximize the connection density between all nodes in the detection results, meaning that there are the closest connections between all nodes in the detection results.

[0166] The process of determining the detection result corresponding to an interactive bipartite graph can also be understood as follows: the interactive bipartite graph is divided into several object detection subgraphs, each with no overlap. That is, each execution of steps S6042 to S6044 yields an object detection subgraph, and each subgraph has no overlap with other object detection subgraphs. When the sum of the density of all obtained object detection subgraphs is maximized, the nodes contained in each of the obtained object detection subgraphs are combined to form a dataset S. d Dataset S d This is the detection result corresponding to the interactive bipartite graph.

[0167] use Figure 9 The method shown can determine the detection result corresponding to each of the N interactive bipartite graphs, resulting in N detection results, which can be denoted as {G1, G2, ..., G...}. n}

[0168] Step S605: From the candidate nodes included in the detection results, identify the abnormal nodes whose risk values ​​reach the set threshold.

[0169] Each node included in a detection result is considered a candidate node. For each candidate node, the following operations are performed: The number of detection results containing that candidate node is determined; this number is called the number of detection results that hit that candidate node. This number is used as the risk value of that candidate node. If the risk value of a candidate node reaches a set threshold, that node is considered an anomalous node. Anomalous nodes may include only merchant nodes, or only consumer nodes; they may also include both merchant and consumer nodes.

[0170] Specifically, assuming a threshold of T is set, for any consumer node in the detection results, if the number of detection results hitting that consumer node is greater than or equal to the set threshold T, it indicates that the probability of that consumer node being abnormal is relatively high. Therefore, that consumer node is added as an abnormal node to the abnormal consumer set u. finalIn the detection results, if the number of detection results hitting that merchant node is greater than or equal to a set threshold T, it indicates that the probability of that merchant node being abnormal is relatively high. Therefore, that merchant node is added as an abnormal node to the abnormal merchant set v. final middle.

[0171] Step S606: Output the account information of the identified abnormal nodes.

[0172] The account corresponding to the abnormal node is the target account. If the abnormal consumer set u determined in step S605... final and abnormal merchant collection v final Neither of them is an empty set, so the set of abnormal consumers u can be represented. final and abnormal merchant collection v final The account information of abnormal nodes is output to the display device, which displays the account information of abnormal consumers and abnormal merchants to the user.

[0173] and Figure 2 The target account identification method shown is based on the same inventive concept. This application also provides a target account identification device, which can be deployed in a server or other electronic device. Since this device corresponds to the target account identification method of this application, and the principle by which this device solves the problem is similar to that of the method, the implementation of this device can refer to the implementation of the above method; repeated details will not be elaborated further.

[0174] Figure 10 This paper presents a schematic diagram of the structure of a target account identification device according to an embodiment of this application. Figure 10 As shown, the target account identification device includes a sampling unit 1001, a detection unit 1002, and an identification unit 1003.

[0175] The sampling unit 1001 is used to generate multiple interaction bipartite graphs based on the resource interaction information between each first type account and each second type account; wherein each interaction bipartite graph contains multiple nodes, and the multiple nodes include at least: first type nodes and second type nodes corresponding to the first type account and the second type account that have resource interaction behavior, respectively.

[0176] The detection unit 1002 is used to detect the connection density between each node in the obtained interactive bipartite graph and obtain the detection results corresponding to each of the multiple interactive bipartite graphs; wherein the connection density between each candidate node in each detection result satisfies the set conditions.

[0177] The identification unit 1003 is used to perform the following operations for each of the obtained candidate nodes: determine the total number of detection results containing one of the candidate nodes, and based on the total number, determine whether the account corresponding to a candidate node is the target account.

[0178] In an optional embodiment, the sampling unit 1001 is specifically used for:

[0179] Get resource interaction information used to record resource exchange behavior between each type 1 account and each type 2 account within a set time period;

[0180] The first type of account in the resource interaction information is taken as the first type of node, and the second type of account in the resource interaction information is taken as the second type of node. Based on the resource exchange behavior between the first type of account and the second type of account, the edges connecting the corresponding first type of node and the corresponding second type of node are generated to obtain the overall resource interaction graph.

[0181] Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected portion of first-type nodes and a portion of second-type nodes, as well as the edges between the selected portion of first-type nodes and a portion of second-type nodes.

[0182] Repeatedly perform the operation of selecting some first-type nodes and some second-type nodes from the overall resource interaction graph to obtain multiple interaction bipartite graphs.

[0183] In an optional embodiment, the sampling unit 1001 is specifically used for:

[0184] The resource interaction graph is transformed into a total adjacency matrix; where the rows in the total adjacency matrix correspond to the second type of nodes in the resource interaction graph, and the columns in the total adjacency matrix correspond to the first type of nodes in the resource interaction graph; the matrix elements in the total adjacency matrix indicate whether there is an edge between the second type of node corresponding to its row and the first type of node corresponding to its column.

[0185] The total adjacency matrix is ​​sampled according to the set sampling rate to obtain the sub-adjacency matrix;

[0186] Based on the second type of nodes corresponding to the rows in the sub-adjacency matrix, the first type of nodes corresponding to the columns in the sub-adjacency matrix, and the matrix elements in the sub-adjacency matrix, the sub-adjacency matrix is ​​transformed into an interactive bipartite graph.

[0187] In an optional embodiment, the sampling unit 1001 is specifically used for:

[0188] Based on the interaction attribute information of the first-class account corresponding to each first-class node, each first-class node in the resource interaction master diagram is divided into the corresponding first-class level; the interaction attribute information includes at least one of the following: total resource interaction amount, resource interaction frequency, or number of resource interactions.

[0189] Based on the interaction attribute information of the second-type account corresponding to each second-type node, the second-type nodes in the resource interaction master diagram are divided into the corresponding second-type levels;

[0190] According to the set sampling rate, a portion of the second-class nodes are selected from the second-class nodes contained in each second-class level, and a portion of the first-class nodes are selected from the first-class nodes contained in each first-class level, according to the sampling rate.

[0191] In an optional embodiment, the detection unit 1002 is specifically used for:

[0192] For each interactive bipartite graph, perform the following operations:

[0193] Get all interactive subgraphs corresponding to one interactive bipartite graph in each interactive bipartite graph; the number of nodes in an interactive subgraph is no greater than the number of nodes in an interactive bipartite graph.

[0194] The density of each interaction subgraph is determined based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node. The density is used to characterize the connection density between nodes in the corresponding interaction subgraph. The degree of a node is the number of edges connected to the node.

[0195] Based on the interaction subgraph with the highest density, determine the detection result corresponding to the interaction bipartite graph.

[0196] In an optional embodiment, the detection unit 1002 is specifically used for:

[0197] For each interactive bipartite graph, perform the following operations:

[0198] Using the interactive bipartite graph as the bipartite graph to be processed, the following operations are performed iteratively until the loop termination condition is met, resulting in multiple candidate detection maps corresponding to the interactive bipartite graph; one iterative process includes:

[0199] Based on the degree of each node in the bipartite graph to be processed, the number of nodes is reduced one by one to obtain multiple subgraphs to be detected; the degree of a node is the number of edges connected to the node.

[0200] The density of each subgraph to be detected is determined based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node. The density is used to characterize the connection density between nodes in the corresponding subgraph to be detected.

[0201] Nodes in the subgraph to be detected that meet the first preset condition of density are added to the candidate detection graph, and the density of the candidate detection graph is determined.

[0202] Remove nodes from the bipartite graph to be processed that meet the first preset density condition to obtain an updated interactive bipartite graph, and use the updated interactive bipartite graph as the bipartite graph to be processed.

[0203] Based on the density of the obtained candidate detection maps, the target detection map is determined from the multiple candidate detection maps, and based on the target detection map, the detection result corresponding to the interaction bipartite map is determined.

[0204] In an optional embodiment, the detection unit 1002 is specifically used for:

[0205] Based on the density of each candidate detection map, determine the second derivative of the density of each candidate detection map.

[0206] Candidate detection maps whose second derivative of density meets the second preset condition are used as target detection maps.

[0207] In an optional embodiment, the identification unit 1003 is specifically used for:

[0208] If the total number is not less than the set threshold, then the account corresponding to a candidate node will be selected as the target account; or,

[0209] The detection result containing a candidate node is used as the target detection result;

[0210] The sum of the reciprocals of the number of nodes contained in each target detection result is used as the detection value corresponding to a candidate node;

[0211] If the detection value corresponding to a candidate node is not less than the set detection threshold, then the account corresponding to the candidate node will be used as the target account.

[0212] Based on the same inventive concept as the above-described method and apparatus embodiments, this application also provides an electronic device. This electronic device may be a server, such as... Figure 1 The server 100 shown can also be a computer or other electronic device. In one embodiment, the structure of the electronic device can be as follows: Figure 11 As shown, it includes a memory 101, a communication module 103, and one or more processors 102.

[0213] The memory 101 is used to store computer programs executed by the processor 102. The memory 101 may mainly include a program storage area and a data storage area. The program storage area may store the operating system and programs required to run instant messaging functions, etc.; the data storage area may store various instant messaging information and operation instruction sets, etc.

[0214] Memory 101 may be volatile memory, such as random-access memory (RAM); memory 101 may also be non-volatile memory, such as read-only memory, flash memory, hard disk drive (HDD), or solid-state drive (SSD); or memory 101 may be any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 101 may be a combination of the above-described memories.

[0215] The processor 102 may include one or more central processing units (CPUs) or digital processing units, etc. The processor 102 is used to implement the aforementioned target account identification method when it invokes a computer program stored in the memory 101.

[0216] The communication module 103 is used to communicate with terminal devices and other electronic devices. If the electronic device is a server, the server can collect resource interaction information of each account through the communication module 103.

[0217] This application embodiment does not limit the specific connection medium between the memory 101, communication module 103, and processor 102 described above. This application embodiment... Figure 11 The memory 101 and the processor 102 are connected via a bus 104, and the bus 104 is in Figure 11 The connections between other components are shown in thick lines only and are not intended to be limiting. Bus 104 can be divided into address bus, data bus, control bus, etc. For ease of illustration, Figure 11 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0218] In some embodiments, the electronic device may further include a display for displaying account information of the target account to be identified.

[0219] In another embodiment, the electronic device can be structured as follows: Figure 12As shown, the memory 101 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 1011 and / or cache memory 1012, and may further include read-only memory (ROM) 1013.

[0220] The memory 101 may also include a program / utility 1015 having a set (at least one) of program modules 1014, including but not limited to: an operating system, one or more application programs, other program modules, and program data, and each or some combination of the examples may include an implementation of a network environment.

[0221] The electronic device can also communicate with one or more external devices 110 (e.g., keyboard, pointing device, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 105. Furthermore, the communication module 103 can be a network adapter, allowing the electronic device to communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via the network adapter. As shown, the communication module 103 communicates with other modules used in the electronic device via bus 104. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0222] In some possible implementations, various aspects of the target account identification method provided in this application can also be implemented as a program product, which includes program code. When the program product is run on a computer device, the program code is used to cause the computer device to perform the steps of the target account identification method according to the various exemplary embodiments of this application described above. For example, the computer device can perform actions such as... Figure 2 The flowchart of the target account identification method in steps S201 to S203 is shown.

[0223] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0224] The program product for the target account identification method according to the embodiments of this application can be a portable compact disc read-only memory (CD-ROM) and include program code, and can run on a computing device. However, the program product of this application is not limited thereto. In this document, the readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0225] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying readable program code. This propagated data signal may take many forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0226] The program code contained on the readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wired, fiber optic, RF, etc., or any suitable combination thereof.

[0227] Program code for performing the operations of this application can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0228] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0229] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0230] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0231] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A target account identification method characterized by, The method includes: Based on the resource interaction information between each type 1 account and each type 2 account, multiple interaction bipartite graphs are generated; each interaction bipartite graph contains multiple nodes, including: type 1 nodes and type 2 nodes corresponding to type 1 accounts and type 2 accounts that have resource interaction behavior, respectively. For each interactive bipartite graph, the interactive bipartite graph is taken as the bipartite graph to be processed, and the following operations are performed in a loop iteration manner until the loop termination condition is met, resulting in multiple candidate detection graphs corresponding to the interactive bipartite graph. Each loop iteration includes: reducing the number of nodes one by one according to the degree of each node in the bipartite graph to be processed, resulting in multiple subgraphs to be detected; the degree of a node is the number of edges connected to that node; and determining the density of each subgraph to be detected based on the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node; the density is used to characterize the corresponding candidate detection graph. The process involves: detecting the connection density between nodes in a subgraph; adding nodes in the subgraph whose density meets a first preset condition to a candidate detection graph and determining the density of the candidate detection graph; removing nodes in the subgraph whose density meets the first preset condition from the bipartite graph to be processed, obtaining an updated interactive bipartite graph, and using the updated interactive bipartite graph as the bipartite graph to be processed; determining a target detection graph from the multiple candidate detection graphs based on their density, and determining the detection result corresponding to the interactive bipartite graph based on the target detection graph; wherein the connection density between candidate nodes in each detection result satisfies a set condition. For each obtained candidate node, perform the following operations: determine the total number of detection results containing the candidate node, and based on the total number, determine whether the account corresponding to the candidate node is the target account.

2. The method of claim 1, wherein, The step of generating multiple interactive bipartite graphs based on the resource interaction information between each type of first-class account and each type of second-class account includes: Get resource interaction information used to record resource exchange behavior between each type 1 account and each type 2 account within a set time period; The first type of account in the resource interaction information is taken as the first type of node, and the second type of account in the resource interaction information is taken as the second type of node. Based on the resource exchange behavior between the first type of account and the second type of account, edges connecting the corresponding first type of node and the corresponding second type of node are generated to obtain the overall resource interaction graph. Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected portion of first-type nodes and the portion of second-type nodes, as well as the edges between the selected portion of first-type nodes and the portion of second-type nodes. Repeatedly perform the operation of selecting a portion of the first type of nodes and a portion of the second type of nodes from the overall resource interaction graph to obtain multiple interaction bipartite graphs.

3. The method of claim 2, wherein, The step of selecting a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generating an interaction bipartite graph based on the selected portion of first-type nodes and the portion of second-type nodes, as well as the edges between the selected portion of first-type nodes and the selected portion of second-type nodes, includes: The resource interaction graph is transformed into a total adjacency matrix; wherein, the rows in the total adjacency matrix correspond to the second type of nodes in the resource interaction graph, and the columns in the total adjacency matrix correspond to the first type of nodes in the resource interaction graph; the matrix elements in the total adjacency matrix indicate whether there is an edge between the second type of node corresponding to its row and the first type of node corresponding to its column. The total adjacency matrix is ​​sampled according to the set sampling rate to obtain the sub-adjacency matrix; Based on the second type of nodes corresponding to the rows in the sub-adjacency matrix, the first type of nodes corresponding to the columns in the sub-adjacency matrix, and the matrix elements in the sub-adjacency matrix, the sub-adjacency matrix is ​​transformed into an interactive bipartite graph.

4. The method of claim 2, wherein, The selection of a portion of the first type of nodes and a portion of the second type of nodes from the overall resource interaction graph includes: Based on the interaction attribute information of the first type of account corresponding to each first type of node, each first type of node in the resource interaction master diagram is divided into the corresponding first type level; the interaction attribute information includes at least one of the following: total resource interaction amount, resource interaction frequency, or number of resource interactions. Based on the interaction attribute information of the second type of account corresponding to each second type of node, each second type of node in the resource interaction master diagram is divided into the corresponding second type level; According to the set sampling rate, a portion of the second-type nodes are selected from the second-type nodes contained in each second-type level, and a portion of the first-type nodes are selected from the first-type nodes contained in each first-type level, according to the same sampling rate.

5. The method of claim 1, wherein, The step of determining the target detection map from the candidate detection maps based on the density of each candidate detection map includes: Based on the density of each candidate detection map, the second derivative of the density of each candidate detection map is determined. Candidate detection maps whose second derivative of density meets the second preset condition are used as the target detection maps.

6. The method of any one of claims 1-5, wherein, Based on the total number, determining whether the account corresponding to the candidate node is the target account includes: If the total number is not less than a set threshold, then the account corresponding to the candidate node is taken as the target account; or, The detection results containing the candidate nodes are used as the target detection results; The sum of the reciprocals of the number of nodes contained in each target detection result is taken as the detection value corresponding to the candidate node; If the detection value corresponding to the candidate node is not less than the set detection threshold, then the account corresponding to the candidate node will be used as the target account.

7. A target account identification apparatus characterized by comprising: The device includes: The sampling unit is used to generate multiple interaction bipartite graphs based on the resource interaction information between each first-type account and each second-type account; wherein each interaction bipartite graph contains multiple nodes, and the multiple nodes include at least: first-type nodes and second-type nodes corresponding to the first-type accounts and second-type accounts that have resource interaction behavior, respectively. The detection unit is used to perform the following operations on each interactive bipartite graph, treating the interactive bipartite graph as a bipartite graph to be processed, and using a loop iterative method until the loop termination condition is met, to obtain multiple candidate detection graphs corresponding to the interactive bipartite graph; wherein, one loop iteration process includes: reducing the number of nodes one by one according to the degree of each node in the bipartite graph to be processed, to obtain multiple subgraphs to be detected; the degree of a node is the number of edges connected to the node; determining the density of each subgraph to be detected according to the number of second-type nodes, the number of first-type nodes, and the degree of each second-type node; the density is used to characterize the... The connection density between nodes in the subgraph to be detected is determined; nodes in the subgraph to be detected whose density meets the first preset condition are added to the candidate detection graph, and the density of the candidate detection graph is determined; nodes in the subgraph to be detected whose density meets the first preset condition are removed from the bipartite graph to be processed, resulting in an updated interactive bipartite graph, and the updated interactive bipartite graph is used as the bipartite graph to be processed; based on the density of each candidate detection graph, a target detection graph is determined from the multiple candidate detection graphs, and based on the target detection graph, the detection result corresponding to the interactive bipartite graph is determined; wherein, the connection density between each candidate node in each detection result satisfies the set condition. The identification unit is configured to perform the following operations for each obtained candidate node: determine the total number of detection results containing the candidate node, and based on the total number, determine whether the account corresponding to the candidate node is the target account.

8. The apparatus according to claim 7, characterized in that, The sampling unit is specifically used for: Get resource interaction information used to record resource exchange behavior between each type 1 account and each type 2 account within a set time period; The first type of account in the resource interaction information is taken as the first type of node, and the second type of account in the resource interaction information is taken as the second type of node. Based on the resource exchange behavior between the first type of account and the second type of account, edges connecting the corresponding first type of node and the corresponding second type of node are generated to obtain the overall resource interaction graph. Select a portion of first-type nodes and a portion of second-type nodes from the overall resource interaction graph, and generate an interaction bipartite graph based on the selected portion of first-type nodes and the portion of second-type nodes, as well as the edges between the selected portion of first-type nodes and the portion of second-type nodes. Repeatedly perform the operation of selecting a portion of the first type of nodes and a portion of the second type of nodes from the overall resource interaction graph to obtain multiple interaction bipartite graphs.

9. The apparatus according to any one of claims 7 to 8, characterized in that, The identification unit is specifically used for: If the total number is not less than a set threshold, then the account corresponding to the candidate node is taken as the target account; or, The detection results containing the candidate nodes are used as the target detection results; The sum of the reciprocals of the number of nodes contained in each target detection result is taken as the detection value corresponding to the candidate node; If the detection value corresponding to the candidate node is not less than the set detection threshold, then the account corresponding to the candidate node will be used as the target account.

10. A computer-readable storage medium storing a computer program therein, characterized in that: When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

11. An electronic device, comprising: It includes a memory and a processor, wherein the memory stores a computer program that can run on the processor, and when the computer program is executed by the processor, it implements the method of any one of claims 1 to 6.