Method, device, equipment and medium for discovering user groups in service applications
By converting order data into user graphs and identifying cheating gangs using community discovery algorithms, the problem of difficult to identify and prevent cheating gangs in Internet service applications in the prior art is solved, and efficient and accurate cheating identification and prevention is achieved.
Patent Information
- Application Number
- CN202011166075.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-10-27
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing technology is difficult to effectively identify and prevent cheating gangs in Internet service applications, resulting in losses of operators' interests. Traditional methods can only identify individual cheating behaviors, cannot identify the entire gang, and are easily avoided.
By converting order data in service applications into user graphs, using graph-based community discovery algorithms (such as LOUVAIN algorithms) to identify cheating gangs based on order relationships between users, aggregating nodes to form sub-graphs, representing user groups, and identifying cheating Gang.
It realizes efficient identification of cheating gangs in service applications, reduces operator losses, improves identification accuracy and efficiency, and is difficult to avoid cheating.
Smart Images

Figure CN112184267B_ABST
Abstract
Description
Technical Field
[0001] Implementations of the present disclosure generally relate to the field of Internet technology, and more specifically, to a method, an apparatus, an electronic device, and a computer-readable storage medium for discovering a user group in a service application. Background Art
[0002] In Internet applications, users are often encouraged to use the services of the Internet application by issuing coupons. For example, in the freight business of transportation applications, the coupons and rewards are very generous. For example, if the order price is 30 yuan, the company's coupon subsidy is 29.99 yuan. Passengers only need to spend 1 cent to complete an order, and the driver can still get normal order income. Therefore, in order to obtain more benefits, many drivers will conduct fake orders and grab a large number of coupons, resulting in damage to the interests of the application service provider. Therefore, it is necessary to carry out necessary risk control to detect cheating users and gangs and prevent the interests of the company from being infringed. Summary of the invention
[0003] An embodiment of the present disclosure provides a technical solution for discovering user groups in a service application.
[0004] In a first aspect of the present disclosure, a method for discovering user groups in a service application is provided, comprising: generating a user graph based on multiple orders placed by multiple users using the service application, wherein nodes of the user graph represent users, two nodes of an edge of the user graph represent a service initiator user and a service recipient user of an order, respectively, and the weight of the edge represents the number of orders generated between corresponding users; and aggregating the nodes in the user graph according to the weights of the edges to obtain one or more subgraphs of the graph, each subgraph representing a user group.
[0005] In a second aspect of the present disclosure, a device for discovering user groups in a service application is provided, comprising: a user graph generating unit, configured to generate a user graph based on multiple orders of multiple users using the service application, wherein the nodes of the user graph represent users, the two nodes of the edge of the user graph represent the service initiator user and the service recipient user of the order respectively, and the weight of the edge represents the number of orders generated between the corresponding users; and an aggregation unit, configured to aggregate the nodes in the user graph according to the weight of the edge to obtain one or more subgraphs of the graph, each subgraph representing a user group.
[0006] In a third aspect of the present disclosure, an electronic device is provided, comprising: a processor and a memory; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method according to the first aspect of the present disclosure.
[0007] In a fourth aspect of the present disclosure, a computer-readable storage medium is provided, on which one or more computer instructions are stored, wherein the one or more computer instructions are executed by a processor to implement the method according to the first aspect of the present disclosure.
[0008] The embodiment of the present disclosure converts the orders in the service application into a graph, and uses an aggregation algorithm based on the graph to find the cheating gang members therein, thereby realizing an efficient anti-cheating technology. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] The features, advantages and other aspects of various implementations of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings, which illustrate several implementations of the present disclosure in an exemplary and non-limiting manner. In the accompanying drawings:
[0010] Figure 1 An example flow chart of a method for discovering a user group in a service application according to an embodiment of the present disclosure is shown;
[0011] Figure 2A shows example fields of a user information table according to an embodiment of the present disclosure;
[0012] Figure 2B shows example fields of an order details table according to an embodiment of the present disclosure;
[0013] Figure 3 An exemplary block diagram showing an apparatus for discovering a user group in a service application according to an embodiment of the present disclosure; and
[0014] Figure 4 A block diagram of a computing device according to an exemplary implementation of the present disclosure is schematically shown. DETAILED DESCRIPTION
[0015] The preferred implementations of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the preferred implementations of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the implementations set forth herein. On the contrary, these implementations are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.
[0016] The term "order" as used in this article refers to a service formed by at least two registered users of an Internet-based service application through the use of the service application by sending and receiving service requests. The service includes but is not limited to travel services for passengers and drivers, and trading services for purchasing physical or virtual goods.
[0017] The term "cheating" as used in this article refers to any behavior performed by users of Internet-based service applications in order to obtain improper benefits, including but not limited to generating improper orders, generating a large number of orders, and any other behavior to obtain improper benefits.
[0018] As used herein, the term "including" and its variations mean open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "based at least in part on". The terms "an example implementation" and "an implementation" mean "at least one example implementation". The term "another implementation" means "at least one additional implementation". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.
[0019] In order to promote their services, operators of Internet applications often issue coupons or rewards to users. For example, in the promotion of the freight business of transportation applications, the coupons and rewards are very generous. For example, if the order price is 30 yuan, the company coupon subsidy is 29.99 yuan. Passengers only need to spend 1 cent to complete an order, and the driver can still get normal order income. Therefore, in order to obtain more benefits, many drivers will engage in order-brushing behavior to grab company coupons. In addition, freight drivers may have multiple passenger accounts while having a driver account, or privately cooperate with multiple passengers to issue orders with passenger accounts when there is no freight demand. After the driver grabs the order, he can start order-brushing behavior. This scenario has its own scenario characteristics. For example, passengers and drivers are in a many-to-many relationship, showing gang characteristics, and even some drivers and passengers are actually in a same-person relationship. For example, drivers and passengers have the same mobile phone number, mobile device number, WeChat account, Alipay account, etc.
[0020] Existing anti-cheating technologies are generally implemented through real-person information and thresholds for the number of fake orders. Real-person information includes mobile phone numbers, mobile device numbers (for example, the IMEI of mobile phones), WeChat accounts, WeChat payment accounts, Alipay accounts, Alipay payment accounts, ID cards, etc. In general business lines, restrictions are imposed on the same ID card, which does not allow drivers and passengers to place orders at the same time, but other real-person information can only be judged after the fact. If driver A and passenger B share the same Alipay account, and driver A and passenger B frequently place orders, they are deemed to have a high probability of fake orders. Secondly, the existing scenarios mainly use the threshold for the number of fake orders to identify cheating, which is achieved by stipulating the threshold for the number of orders placed by the same pair of drivers and passengers. If the number of orders completed by driver A and passenger B in one day exceeds 10 times, they are deemed to have a high probability of fake orders.
[0021] However, the above method can only identify and process individual cases, but cannot obtain information about the entire group, resulting in low execution efficiency. As for the method of setting a threshold for the number of fake orders, the cheating group can easily find out the rules of the strategy and evade the strategy by trying to explore the threshold limit, thereby limiting the number of fake orders within the threshold.
[0022] In view of the above, an efficient and non-avoidable method is needed to identify cheating groups to reduce the losses of application operators. The present disclosure proposes a method for discovering user groups in service applications, which can identify groups that implement cheating behaviors in service applications. According to an embodiment of the present disclosure, based on the order relationship between drivers and passengers and the number of orders between them, a graph-based community discovery algorithm can be used to identify cheating groups.
[0023] Figure 1 An example flow chart of a method 100 for discovering user groups in a service application according to an embodiment of the present disclosure is shown, comprising: at step 110, a user graph is generated based on multiple orders of multiple users using the service application, the nodes of the user graph represent users, the two nodes of the edge of the user graph represent the service initiator user and the service recipient user of the order respectively, and the weight of the edge represents the number of orders generated between the corresponding users. The method 100 also includes: at step 120, the nodes in the user graph are aggregated according to the weight of the edge to obtain one or more subgraphs of the user graph, each subgraph represents a user group.
[0024] The following describes in detail step 110 of generating a user graph. Before executing step 110, a user information table and an order detail table may be obtained from a database, and the obtained tables may be preprocessed by data cleaning, data screening, data dimension reduction, etc. to generate an order data set for constructing a user graph.
[0025] Data preprocessing includes but is not limited to removing duplicate data, null value data, abnormal data, etc. from the table, and selecting fields required by the algorithm from the table. Since data tables often contain a lot of redundant information, it is necessary to clean the data and select the required fields to reduce data redundancy and improve the efficiency of the algorithm. The information of these tables themselves and the cross-matching information between these tables will be used to determine whether the recommender has a high suspicion of cheating.
[0026] Figure 2A 2 shows example fields of the user information table 210 according to an embodiment of the present disclosure. Figure 2AAs shown, after data preprocessing, the user information table 210 may include information associated with the user's real identity, including but not limited to: user ID, real name, ID number, mobile phone number, encrypted mobile phone number (e.g., md5 encryption), mobile device number (e.g., IMEI of mobile phone), social network application ID (e.g., WeChat account, etc.), payment application ID (e.g., Alipay account, bank debit card or credit card account, etc.), biometrics (e.g., encrypted face recognition features, fingerprint features, etc.), or other information associated with the user's real identity information. It should be noted that in the service application, in addition to the user ID that uniquely identifies the user, other attributes or fields in the user information table 210 can have multiple values. For example, a user can have multiple mobile phone numbers (and associated encrypted mobile phone numbers), multiple mobile device numbers, multiple social network user IDs, multiple payment application IDs, etc. associated with him. In addition, in some embodiments, the user ID can be divided into multiple types, for example, a passenger ID and a driver ID, the passenger ID can only be used to initiate a service request, and the driver ID can only be used to accept a service request. Moreover, in some embodiments, both the passenger ID and the driver ID can be registered using identifiers such as real name, ID number, mobile phone number, etc., so that one identifier may correspond to multiple user IDs within the application.
[0027] like Figure 2B As shown, after data preprocessing, the order details table 220 may include but is not limited to: order ID, passenger ID, passenger information (e.g., encrypted mobile phone number), driver ID, driver information (e.g., encrypted mobile phone number), order status, start / finish time, order duration, starting location, end location, order mileage, order fee, coupon amount, etc. The order ID is the primary key of the order details table 220 and is used to uniquely identify the order. In a transportation application, the passenger ID and the driver ID can represent two types of user IDs in the travel order formed between users in the application. For example, a user in the application can initiate a car request, and in response to the user request being accepted by another user to generate an order, the user ID that initiates the car request can be recorded as the passenger ID, and the user ID that accepts the request can be recorded as the driver ID, and vice versa. It can be seen from this that through the user ID information used to uniquely identify the user in the application, information query between the cross-user information table 210 and the order details table 220 can be realized. Other information in the order details table 220, such as order status, start / completion time, order duration, starting location, end location, order mileage, order fee, coupon amount and other information, can be used to check whether the order is an abnormal order, thereby determining whether the order is an abnormal order.
[0028] According to the embodiment of the present disclosure, the order details table 220 may be filtered. For example, in the application scenario of the present invention, since the purpose of cheating is to obtain order coupons or rewards for cashing out, the orders in the filtered order details table may be limited to have a value greater than or equal to a preset threshold C. k In addition, the order status can be filtered as completed order. In one embodiment, orders within a period of time can be selected for cheating analysis. For example, the start or completion time of the order can be limited to a specified time window, such as 10 days, but not limited thereto, so that orders within these 10 days can be obtained.
[0029] According to an embodiment of the present disclosure, user information that is different between the service requester and the service receiver that generates the order can be used as the primary key. For example, the primary key of the driver and the passenger can be represented by a mobile phone number encrypted with md5. Alternatively, other real identity information, such as an ID number, can also be used. Since in the scenario where an order is generated and the driver ID and the passenger ID are two different types of IDs (i.e., the passenger ID is only used to initiate an order, and the driver ID is only used to accept an order), the passenger and the driver may exchange roles, such as A registering the driver account D1 with a mobile phone number and registering the passenger account P1 with the same mobile phone number; B registering the driver account D2 with a mobile phone number and registering the passenger account P2 with the same mobile phone number. (P1, D2), (P2, D1) are actually joint orders of real identities (A, B). Therefore, in the identification of cheating gangs, it may be more reasonable to use real identity information such as mobile phone numbers as the primary key, which increases the accuracy of gang identification.
[0030] In one embodiment, when the order details table 220 includes the encrypted mobile phone number of the user, an order data set for constructing a user graph can be generated from the order details table 220, and each element in the data set has a form such as (driver D, passenger P, order quantity), wherein the driver D and the passenger P can be the encrypted mobile phone numbers of the corresponding users as unique identifiers. Alternatively, when the order details table 220 does not include the encrypted mobile phone number of the user, the driver ID or the passenger ID can be used to query the corresponding encrypted mobile phone number in the user information table 210, thereby forming an order data set for constructing a user graph, which includes the elements of (driver D, passenger P, order quantity).
[0031] In addition, during the statistical process, there may be a real need for merchants to issue a large number of orders, and there may be a normal phenomenon that merchants and drivers do have a high number of orders. In order to prevent merchants from mistakenly believing that they are cheating gangs, the merchant list can be excluded in advance during the statistical process.
[0032] In addition, according to an embodiment of the present disclosure, when forming an order data set, a threshold T may also be set. c, so that the order data set (driver D, passenger P, order quantity N) used to construct the user graph only includes the order quantity greater than the threshold T c The driver and passenger data set, threshold T c The number may be 3, but is not limited thereto. Thus, excessive noise in the user graph can be avoided, thereby improving recognition accuracy.
[0033] According to an embodiment of the present disclosure, the above-mentioned order data set (driver D, passenger P, order quantity N) can be used to generate a user graph. For example, for each element in the order data set, driver D and passenger P can be mapped to nodes on the user graph, and the order quantity is mapped to the edge between the two nodes on the user graph as the weight of the edge. According to an embodiment of the present disclosure, if there is no edge between nodes on the user graph, it means that no order has been formed between the two nodes; the weight of the edge between nodes on the user graph represents the number of orders formed between the two users. It should be noted that in the scenario of the present invention, the driver and the passenger may exchange roles, so the driver and the passenger are not distinguished here, and the md5 encrypted mobile phone number is used as the node in the user graph. In this case, the elements in which the driver and the passenger exchange roles in the order data set can be superimposed, that is, the weight of the edge is the sum of the order quantity N in the two elements. For example, the order data set has element E1 (a, b, order quantity N1), that is, b as a passenger and a as a driver have N1 orders), element E2 (b, a, order quantity N2), that is, a as a passenger and b as a driver have N2 orders, where a and b are encrypted mobile phone numbers. At this time, the weight of the edge between node a and node b can be N1+N2. In other words, the weight of the edge represents the sum of the number of orders formed by the corresponding user as either the service initiator or the service receiver. Thus, the generated user graph can be an undirected graph.
[0034] As described in step 120 of method 100, the nodes in the user graph are aggregated according to the weights of the edges to obtain one or more subgraphs of the user graph, each subgraph representing a user group. According to an embodiment of the present disclosure, a cheating gang identification model can be used to perform graph algorithm clustering on the nodes, and the clustering result can be one or more subgraphs of the user graph, each subgraph representing a user group, that is, a gang identification result. In one embodiment, the clustering result can be a set of encrypted mobile phone numbers; alternatively, the clustering result can be in another form of {D1, D2, D3, P1, P2, P3, P4, P5, P6, P7}, that is, the gang members include multiple passengers and drivers. It can be understood that in combination Figure 2A The user information table 220 shown can be used to query the corresponding passenger ID and driver ID using the encrypted mobile phone number.
[0035] According to an embodiment of the present disclosure, the graph clustering algorithm may be, for example, a community discovery algorithm LOUVAIN algorithm, but is not limited thereto. The LOUVAIN algorithm may use the modularity of the community as the objective function Q, and maximize the modularity of the entire community network, that is, to make the entire community network present a module cluster, that is, a structure of one or more subgraphs (also referred to as communities). The steps of the LOUVAIN algorithm are briefly described below:
[0036] 1) Treat each node in the graph as an independent community. Initially, the number of communities is the same as the number of nodes.
[0037] 2) For each node i, try to assign node i to the community where each of its neighboring nodes is located in turn, calculate the modularity change Delta Q before and after the assignment, and record the neighboring node with the largest Delta Q. If maxDelta Q>0, assign node i to the community where the neighboring node with the largest Delta Q is located, otherwise it remains unchanged;
[0038] 3) Repeat 2) until the communities to which all nodes belong no longer change;
[0039] 4) Compress the graph, compress all nodes in the same community into a new node, convert the edge weights between nodes in the community into the ring weights of the new node, and convert the edge weights between communities into edge weights between new nodes;
[0040] 5) Repeat 1) until the modularity of the entire graph no longer changes.
[0041] According to an embodiment of the present disclosure, modularity Q can be calculated based at least in part on the sum of edge weights within one or more subgraphs and the sum of edge weights between one or more subgraphs. Thus, by maximizing modularity Q, adjacent nodes with larger edge weights can be clustered into the same user group.
[0042] It should be noted that for the user graph clustering algorithm, it is necessary to collect the training set data required for training and the input data required after the cheating group model implemented by the present invention is launched. In addition to the order details table obtained from the database, the training data set also includes a verification step, which verifies whether the driver and passengers have indeed joined forces to cheat. Since only groups with a large number of group members are meaningful for fraudulent order identification, after obtaining the user group of the group identification result, it is possible to derive the number of nodes in the user group that is greater than or equal to the preset threshold N. g A user group with a number of users below the threshold N will be identified as a cheating group. Alternatively, it is also possible to derive g user base.
[0043] According to an embodiment of the present disclosure, verification can be an offline verification, including verifying whether there is a many-to-many order relationship between drivers and passengers in the user group, that is, whether it is a gang relationship. For example, the number of edges and nodes in the user group can be calculated, and based on the relationship between the number of edges and the number of nodes in the user group, it is determined whether there is a many-to-many order relationship. It can be understood that in the subgraph of the user group, for a given number of nodes, the more edges there are, the greater the possibility of the existence of a many-to-many order relationship. A function related to the number of nodes can be set, and the number of edges of the user group can be compared with the function for offline verification of whether there is a many-to-many order relationship among the user group.
[0044] In addition, offline verification can also include verifying whether the orders between users in the user group are abnormal orders (i.e., whether the gang members cheat). Abnormal orders have any one or more of the following: the service initiator and the service receiver have the same social attributes, the order start and end points overlap or are close, the order duration is too short, the order mileage is too short, the order is issued in different places, and multiple orders with the start and end points clustered in the subgraph (user group). For example, you can refer to Figure 2B The order details table 220 is used to verify abnormal orders.
[0045] Through verification, the accuracy, precision, and recall of the user group obtained by the current cheating group model can be obtained as the verification result. According to the embodiments of the present disclosure, the accuracy, precision, and recall of the model can be used to further adjust the model parameters, improve the accuracy and precision of the model, and reduce the recall rate. For example, in the LOUVAIN algorithm, the size of the community can be changed by adjusting the resolution parameter, and the threshold number of members N of the user group can also be adjusted. g , to get a more accurate picture of the cheating gangs.
[0046] From the above, by iterating and optimizing the model multiple times, a model for identifying cheating gangs that can be put online is obtained. For example, after collecting the input data required for going online, such as the order details list within 10 days, the corresponding order data set (driver D, passenger P, order volume N) is input into the trained gang identification model. Figure 1 The method 100 is used to discover user groups in the service application. The identified user groups may include a set of encrypted mobile phone numbers and a set of user IDs (including passenger IDs and driver IDs) of the service application.
[0047] Figure 3An example block diagram of an apparatus 300 for discovering user groups in a service application according to an embodiment of the present disclosure is shown. The apparatus 300 includes: a user graph generation unit 310, configured to generate a user graph based on multiple orders of multiple users using the service application, wherein the nodes of the user graph represent users, the two nodes of the edge of the user graph represent the service initiator user and the service recipient user of the order respectively, and the weight of the edge represents the number of orders generated between the corresponding users; and an aggregation unit 320, configured to aggregate the nodes in the user graph according to the weight of the edge to obtain one or more subgraphs of the user graph, each subgraph representing a user group.
[0048] According to an embodiment of the present disclosure, the nodes of the user graph may be identified by information associated with the user's real identity, such as a mobile phone number (unencrypted or encrypted), a mobile device number, a social network application ID, a payment application ID, biometrics, etc. As described above, in the identification of cheating gangs, it may be more reasonable to use such information associated with the real identity as the primary key, which increases the accuracy of gang identification.
[0049] According to an embodiment of the present disclosure, the weight of an edge may represent the sum of the number of orders formed by the corresponding user as either the service initiator or the service receiver. In some cases, the roles of the service initiator and the service receiver may be interchanged, so their order data may be added together as the weight of the edge, thereby generating an undirected graph.
[0050] According to an embodiment of the present disclosure, the weight of the edge may be at least greater than or equal to the first threshold. In some cases, nodes with fewer orders may be filtered out to remove noise, thereby improving execution efficiency and accuracy.
[0051] According to an embodiment of the present disclosure, the number of nodes included in each subgraph is greater than or equal to the second threshold. Since only groups with a large number of group members are meaningful for fraudulent order identification, after obtaining the user group of the group identification result, the user group with the number of nodes in the user group greater than or equal to the preset threshold can be derived and identified as a cheating group.
[0052] According to an embodiment of the present disclosure, aggregation may include: initializing so that each node belongs to a subgraph that it constitutes; iteratively traversing each node in the user graph, and determining whether to merge the current node into other subgraphs to which nodes connected to it belong by maximizing the modularity of the user graph, wherein the modularity is calculated at least in part based on the sum of the weights of the edges within one or more subgraphs and the sum of the weights of the edges between one or more subgraphs.
[0053] According to an embodiment of the present disclosure, the device 300 may further include a verification unit 330, which is configured to verify whether the users in the user group have abnormal orders when using the application to adjust the aggregated parameters based on the verification results. For example, by verifying whether the users in the user group have abnormal orders, the accuracy, precision, and recall of the aggregation can be determined, thereby adjusting the aggregated parameters, that is, the model parameters can be further adjusted through verification to improve the accuracy and precision of the model and reduce the recall rate.
[0054] According to an embodiment of the present disclosure, abnormal orders may include any one or more of the following: orders in which the service initiator user and the service recipient user have the same social attributes, orders with overlapping or close starting and ending points, orders with too short a duration, orders with too short a mileage, and multiple orders with clustered starting and ending points within the user group.
[0055] The units included in the device 300 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units can be implemented using software and / or firmware, such as machine executable instructions stored on a storage medium. In addition to or as an alternative to machine executable instructions, some or all of the units in the device 300 can be implemented at least in part by one or more hardware logic components. As an example and not limitation, exemplary types of hardware logic components that can be used include field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.
[0056] Figure 4 4 shows a block diagram of a computing device / server 400 in which one or more embodiments of the present disclosure may be implemented. It should be understood that Figure 4 The computing device / server 400 shown is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein.
[0057] like Figure 4 As shown, computing device / server 400 is in the form of a general-purpose computing device. Components of computing device / server 400 may include, but are not limited to, one or more processors or processing units 410, memory 420, storage device 430, one or more communication units 440, one or more input devices 450, and one or more output devices 460. Processing unit 410 may be an actual or virtual processor and is capable of performing various processes according to a program stored in memory 420. In a multi-processor system, multiple processing units execute computer executable instructions in parallel to increase the parallel processing capabilities of computing device / server 400.
[0058] The computing device / server 400 typically includes a plurality of computer storage media. Such media may be any available media accessible to the computing device / server 400, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory 420 may be a volatile memory (e.g., registers, caches, random access memory (RAM)), a non-volatile memory (e.g., a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device 430 may be a removable or non-removable medium, and may include a machine-readable medium, such as a flash drive, a disk, or any other medium, which may be capable of being used to store information and / or data (e.g., training data for training) and may be accessed within the computing device / server 400.
[0059] The computing device / server 400 may further include additional removable / non-removable, volatile / non-volatile storage media. Figure 4 As shown in , a disk drive for reading or writing from a removable, non-volatile disk (e.g., a "floppy disk") and an optical drive for reading or writing from a removable, non-volatile optical disk may be provided. In these cases, each drive may be connected to the bus (not shown) by one or more data media interfaces. The memory 420 may include a computer program product 425 having one or more program modules that are configured to perform various methods or actions of various embodiments of the present disclosure.
[0060] The communication unit 440 enables communication with other computing devices via a communication medium. Additionally, the functions of the components of the computing device / server 400 can be implemented in a single computing cluster or multiple computing machines that can communicate via a communication connection. Thus, the computing device / server 400 can operate in a networked environment using a logical connection with one or more other servers, a network personal computer (PC), or another network node.
[0061] Input device 450 may be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 460 may be one or more output devices, such as a display, speaker, printer, etc. Computing device / server 400 may also communicate with one or more external devices (not shown) as needed, such as storage devices, display devices, etc., with one or more devices that enable a user to interact with computing device / server 400, or with any device that enables computing device / server 700 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.) through communication unit 440. Such communication may be performed via an input / output (I / O) interface (not shown).
[0062] According to an exemplary implementation of the present disclosure, a computer-readable storage medium is provided, on which one or more computer instructions are stored, wherein the one or more computer instructions are executed by a processor to implement the method described above.
[0063] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products implemented according to the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.
[0064] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.
[0065] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.
[0066] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple implementations of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some implementations as replacements, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be realized by a special hardware-based system that performs the function or action of the specification, or can be realized by a combination of special hardware and computer instructions.
[0067] The above descriptions of various implementations of the present disclosure are exemplary, non-exhaustive, and not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The selection of terms used herein is intended to best explain the principles of the implementations, practical applications, or improvements to the technology in the marketplace, or to enable other persons of ordinary skill in the art to understand the implementations disclosed herein.
Claims
1. A method for discovering a user group in a service application, comprising: Performing data preprocessing on the user information and the order information to generate an order data set for constructing a user graph, wherein the order data set uses the user's real identity information that makes the service requester and the service provider who generate the order different as a primary key; A user graph is generated based on multiple orders of multiple users using the service application, wherein the nodes of the user graph represent users, the two nodes of the edge of the user graph represent the service initiator user and the service receiver user of the order respectively, the weight of the edge represents the number of orders generated between the corresponding users, the weight of the edge represents the sum of the number of orders formed by the corresponding user as the service initiator or the service receiver when using the application, and the weight of the edge is at least greater than or equal to a first threshold; as well as aggregating the nodes in the user graph according to the weights of the edges to obtain one or more subgraphs of the user graph, each subgraph representing a user group; The aggregating the nodes in the user graph according to the weights of the edges to obtain one or more subgraphs of the user graph includes: Using a cheating gang identification model to perform graph algorithm clustering on the nodes to obtain a clustering result, wherein the clustering result includes one or more subgraphs of the user graph; The cheating group identification model is trained based on a training data set, and the training data set at least includes an order details table and joint cheating data obtained through verification.
2. The method according to claim 1, wherein: The nodes of the user graph are identified by information associated with the real identity of the user.
3. The method according to claim 1, wherein: The number of nodes included in each subgraph is greater than or equal to a second threshold.
4. The method according to claim 1, wherein: The polymerization includes: Initialize so that each node belongs to its own subgraph; Iteratively traverse each node in the user graph, and determine whether to merge the current node into other subgraphs to which nodes connected to the current node belong by maximizing the modularity of the user graph, wherein the modularity is calculated at least in part based on the sum of weights of edges within the one or more subgraphs and the sum of weights of edges between the one or more subgraphs.
5. The method according to claim 1, further comprising: Verify whether users in the user group place abnormal orders when using the application, so as to adjust the aggregation parameters based on the verification result.
6. The method according to claim 5, wherein: The application is a vehicle service application, and the abnormal orders include any one or more of the following: orders in which the service initiator user and the service recipient user have the same social attributes, orders with overlapping or close starting and ending points, orders with too short a duration, orders with too short a mileage, and multiple orders with clustered starting and ending points within the user group.
7. A device for discovering a user group in a service application, comprising: A user graph generating unit is configured to generate a user graph based on multiple orders of multiple users using the service application, wherein the nodes of the user graph represent users, the two nodes of the edge of the user graph represent the service initiator user and the service receiver user of the order respectively, the weight of the edge represents the number of orders generated between the corresponding users, the weight of the edge represents the sum of the number of orders formed by the corresponding user as the service initiator or the service receiver when using the application, and the weight of the edge is at least greater than or equal to a first threshold; as well as an aggregation unit, configured to aggregate the nodes in the user graph according to the weights of the edges to obtain one or more subgraphs of the user graph, each subgraph representing a user group; The device of the user group is further configured to perform data preprocessing on the user information and the order information to generate an order data set for constructing a user graph, wherein the order data set uses the user's real identity information that makes the service requester and the service provider who generate the order different as the primary key; The polymeric unit is specifically used for: Using a cheating gang identification model to perform graph algorithm clustering on the nodes to obtain a clustering result, wherein the clustering result includes one or more subgraphs of the user graph; The cheating group identification model is trained based on a training data set, and the training data set at least includes an order details table and joint cheating data obtained through verification.
8. The device according to claim 7, wherein: The nodes of the user graph are identified by information associated with the real identity of the user.
9. The device according to claim 7, wherein: The number of nodes included in each subgraph is greater than or equal to a second threshold.
10. The device according to claim 7, wherein: The aggregation unit is further configured to: Initialize so that each node belongs to its own subgraph; and Iteratively traverse each node in the user graph, and determine whether to merge the current node into other subgraphs to which nodes connected to the current node belong by maximizing the modularity of the user graph, wherein the modularity is calculated at least in part based on the sum of weights of edges within the one or more subgraphs and the sum of weights of edges between the one or more subgraphs.
11. The apparatus according to claim 7, further comprising: The verification unit is configured to verify whether users in the user group have placed abnormal orders when using the application, so as to adjust the aggregation parameters based on the verification result.
12. The device according to claim 11, wherein The abnormal orders include any one or more of the following: orders in which the service initiator user and the service recipient user have the same social attributes, orders with overlapping or close starting and ending points, orders with too short a duration, orders with too short a mileage, and multiple orders with clustered starting and ending points within the user group.
13. An electronic device, comprising: processor; as well as A memory coupled to the processor and containing instructions stored thereon, which, when executed by the processor, cause the electronic device to perform the method according to any one of claims 1 to 6.
14. A computer-readable storage medium having one or more computer instructions stored thereon, wherein the one or more computer instructions are executed by a processor to implement the method according to any one of claims 1 to 6.
