User group mining method, user group mining device, storage medium and electronic equipment

Abstract

The embodiment of the invention relates to a user group mining method and device, a storage medium and electronic equipment, and relates to the technical field of Internet. The user group mining method comprises the following steps: constructing a bipartite graph according to a service relation between a resource creation user and a resource consumption user; the resource creation users are associated with creation resources, and the bipartite graph comprises resource creation user nodes used for representing each resource creation user, resource consumption user nodes used for representing each resource consumption user, and connection edges between the resource creation user nodes with business connection and the resource consumption user nodes; determining graph parameters associated with the bigraph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node and the edge weight of each connecting edge; and carrying out illegal user group mining based on the graph parameters and the bipartite graph to obtain an illegal user group mining result. According to the invention, the illegal user group identification accuracy can be improved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of Internet technology, and more specifically, relate to a user group mining method, a user group mining device, a computer-readable storage medium, and an electronic device. Background Art

[0002] This section is intended to provide a background or context to embodiments of the present disclosure, and no description herein is admitted to be prior art by inclusion in this section.

[0003] UGC is the abbreviation of "User-Generated Content", which means user-generated content. The content forms include but are not limited to songs, short videos, long videos, etc. Users create UGC content, and the platform measures the quality of the work through dimensions such as the number of views, and then distributes corresponding benefits to users. In this process, some illegal activities such as black market volume increase are prone to occur. Summary of the invention

[0004] In related technologies, the following solutions are generally used to solve the problem of excessive broadcasting in UGC scenarios:

[0005] (1) Detect and identify spammers through statistical type indicators, data mart rules, and other methods, and then aggregate them into the creator dimension for disposal;

[0006] (2) Constructing related groups with users who broadcast excessively as nodes through dimensions such as playback devices and IP (internet protocol address). After identifying the groups of users who broadcast excessively, they are aggregated to the creator dimension for disposal;

[0007] (3) Construct the associated groups of the two nodes of the broadcaster and the creator through static dimensions such as the playback device and IP address. After identifying the two groups of the broadcaster and the creator, deal with the identified creators;

[0008] However, the inventors found that the above solution has the following problems:

[0009] (1) Statistical indicators and data mart rules cannot effectively mine gang associations, and the recall is insufficient and there are omissions. In addition, in the fierce confrontation with the black market, the rules or thresholds need to be constantly adjusted, which has a high maintenance cost.

[0010] (2) The creator dimension information is lost. It is difficult to identify whether the consumption traffic is risky by only using the traditional resource consumer playback device, IP, etc., which may cause accidental damage to the creators of popular resources in the fan group brushing scenario;

[0011] (3) The two-part graph of the broadcaster-creator also only considers the dimensions of the broadcaster’s device and IP address. It is also unable to distinguish whether it is natural traffic or traffic from illegal small accounts, which can easily lead to accidental damage to natural traffic.

[0012] The above shortcomings will also have the following impacts on the platform business:

[0013] (1) Risky traffic is leaked, causing the platform to suffer huge economic losses and also bear high risk control rules dimension costs;

[0014] (2) If normal fan traffic is accidentally damaged, it is easy to cause customer complaints, create adverse social impacts, and affect the platform's reputation;

[0015] (3) Natural traffic is accidentally damaged, which dampens the enthusiasm of users who create resources and affects the creative ecology of the UGC platform.

[0016] Therefore, there is a great need for a user group mining method that can more accurately mine the illegal user groups.

[0017] In this context, embodiments of the present disclosure are intended to provide a user group mining method, a user group mining device, a computer-readable storage medium, and an electronic device.

[0018] According to a first aspect of an embodiment of the present disclosure, a user group mining method is provided, comprising:

[0019] A bipartite graph is constructed based on the business relationship between resource creation users and resource consumption users; the resource creation users are associated with creation resources, and the bipartite graph includes a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes having business relationships;

[0020] Determine the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge; the edge weight of each connecting edge is determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations;

[0021] Based on the graph parameters and the bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

[0022] In an optional embodiment, the graph parameters associated with the bipartite graph include first part parameters and second part parameters;

[0023] The mining of illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results includes:

[0024] Pruning the bipartite graph based on the first part of parameters to obtain a core bipartite graph;

[0025] Based on the second part of parameters and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

[0026] In an optional implementation, the resource consumption user characteristics include the user activity of each resource consumption user, the resource creation user characteristics include the resource popularity of each resource creation user, and the first part of parameters includes the user activity, the resource popularity and the traffic contribution;

[0027] The pruning of the bipartite graph based on the first part of parameters includes:

[0028] Comparing the resource heat associated with each resource creation user node with a preset heat threshold, and performing pruning processing on the resource creation user node when the resource heat is not greater than the preset heat threshold;

[0029] When the resource heat is greater than the preset heat threshold, comparing the user activity associated with each of the resource consuming user nodes with the preset activity threshold;

[0030] When the user activity is not greater than the preset activity threshold, pruning the resource consuming user node;

[0031] When the user activity is greater than the preset activity threshold, comparing the traffic contribution associated with each edge with the preset traffic contribution threshold;

[0032] When the traffic contribution is not greater than a preset traffic contribution threshold, pruning is performed on the edge.

[0033] In an optional implementation, the traffic contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business connection is calculated in the following way:

[0034] Obtaining the total consumption flow of the creation resource of the resource creation user within a preset time period, and obtaining the partial consumption flow of each resource consumption user for the creation resource within the preset time period;

[0035] According to the ratio between each of the partial consumption flows and the total consumption flow, the flow contribution of each resource consuming user to the creation resources of the resource creation users with whom the user has business relations is determined.

[0036] In an optional implementation, the resource heat is calculated in the following way:

[0037] Calculate the average consumption flow of the creation resource of the resource creation user within a preset time period;

[0038] The resource heat is determined according to the average consumption flow.

[0039] In an optional implementation manner, the user activity is calculated in the following manner:

[0040] Obtaining the average consumption flow of the resource consuming user for historical consumption resources within a preset time period;

[0041] The user activity is determined according to the average consumption flow.

[0042] In an optional implementation, the resource consumption user characteristics further include resource consumption user attributes, the resource creation user characteristics further include resource creation user attributes, and the second part of parameters includes the resource consumption user attributes and the resource creation user attributes;

[0043] The mining of illegal user groups based on the second part of parameters and the core bipartite graph to obtain illegal user group mining results includes:

[0044] Based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

[0045] In an optional implementation, the mining of illegal user groups based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph to obtain illegal user group mining results includes:

[0046] Performing user group mining based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph through a preset graph mining algorithm to obtain illegal user group mining results;

[0047] Among them, the preset graph mining algorithm includes at least one of the following: a community discovery algorithm and a deep learning algorithm.

[0048] According to a second aspect of the embodiments of the present disclosure, there is provided a user group mining device, comprising:

[0049] A graph construction module, for constructing a bipartite graph according to the business relationship between resource creation users and resource consumption users; the resource creation users are associated with creation resources, and the bipartite graph includes a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes having business relationships;

[0050] A graph parameter determination module, for determining the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge; the edge weight of each connecting edge is determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations;

[0051] The user group mining module is used to mine illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results.

[0052] According to a third aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the user group mining method described in the first aspect is implemented.

[0053] According to a fourth aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the user group mining method described in the first aspect above by executing the executable instructions.

[0054] According to the user group mining method, user group mining device, computer-readable storage medium and electronic device implemented in the present disclosure, on the one hand, by constructing a bipartite graph between resource creation users and resource consumption users, and according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node and the edge weight of each link, the graph parameters associated with the bipartite graph are determined, so that each node and link in the bipartite graph is associated with rich and detailed multi-dimensional feature information, thereby facilitating a more comprehensive understanding of user behavior and providing a basis for the subsequent identification of illegal user groups; further, the weight of each link reflects the resource consumption user's influence on the resource creation user. Traffic contribution, which can more accurately quantify the interaction intensity and relationship closeness between users, more accurately locate the core focus of illegal behaviors such as broadcasting, and further reduce the subsequent probability of misidentification; on the other hand, based on the graph parameters and bipartite graph, the illegal user group mining is carried out to obtain the illegal user group mining results, which can solve the problem of inaccurate identification results caused by identifying the broadcasting user group only based on static features or set fixed rules in related technologies, and obtain more accurate illegal user group mining results. In addition, on the basis of accurately identifying the illegal user group, it is also helpful to efficiently deal with illegal users, improve the platform's governance capabilities and user experience, and enhance user trust and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present disclosure are shown in an exemplary and non-limiting manner, in which:

[0056] Figure 1 A schematic diagram showing a flow chart of a user group mining method in an embodiment of the present disclosure;

[0057] Figure 2 A flow chart showing how to determine the resource popularity of each resource creation user in an embodiment of the present disclosure;

[0058] Figure 3 A schematic diagram showing a flow chart of how to determine the user activity of each resource consuming user in an embodiment of the present disclosure;

[0059] Figure 4 A flow chart showing how to determine the flow contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business relationship in the embodiment of the present disclosure;

[0060] Figure 5 A schematic diagram showing another process of obtaining illegal user group mining results in an embodiment of the present disclosure;

[0061] Figure 6 A schematic diagram showing a process of pruning a bipartite graph based on the first part of parameters to obtain a core bipartite graph in an embodiment of the present disclosure;

[0062] Figure 7 A schematic diagram showing the overall process of how to prune a bipartite graph to obtain a core bipartite graph in an embodiment of the present disclosure;

[0063] Figure 8 A schematic diagram showing the overall process of the user group mining method in an embodiment of the present disclosure is shown;

[0064] Fig. 9 A schematic diagram of a user group mining device according to an embodiment of the present disclosure is shown;

[0065] Fig.10 A structural diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0066] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0067] The principles and spirit of the present disclosure will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present disclosure, and are not intended to limit the scope of the present disclosure in any way. On the contrary, these embodiments 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.

[0068] Those skilled in the art will appreciate that the embodiments of the present disclosure may be implemented as a system, device, apparatus, method or computer program product. Therefore, the present disclosure may be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0069] According to an embodiment of the present disclosure, a user group mining method, a user group mining device, a computer-readable storage medium, and an electronic device are provided.

[0070] In this document, any number of elements in the drawings is used for illustration rather than limitation, and any naming is used only for distinction and does not have any limiting meaning.

[0071] The principle and spirit of the present disclosure are explained in detail below with reference to several representative embodiments of the present disclosure. SUMMARY OF THE INVENTION

[0073] The inventors have discovered that existing solutions for identifying illegal user groups are not accurate enough, and there is a tendency for missed identification or misidentification, which can easily lead to accidental damage to normal creative traffic, dampen the enthusiasm of creators, and thus affect the reputation of the platform's operations.

[0074] In view of the above content, the basic idea of ​​the present disclosure is to provide a user group mining method, a user group mining device, a computer-readable storage medium and an electronic device, by constructing a bipartite graph between resource creation users and resource consumption users, and determining the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node and the edge weight of each edge, so that each node and edge in the bipartite graph is associated with rich and detailed multi-dimensional feature information, thereby facilitating a more comprehensive understanding of user behavior and providing a basis for the subsequent identification of illegal user groups; further, the weight of each edge reflects the resource consumption user's influence on the resource creation. The traffic contribution of users can be calculated more accurately, which can more precisely quantify the interaction intensity and relationship closeness between users, more accurately locate the core concerns of illegal behaviors such as broadcasting, and further reduce the subsequent probability of misidentification. On the other hand, illegal user groups are mined based on graph parameters and bipartite graphs to obtain illegal user group mining results. This can solve the problem of inaccurate identification results caused by identifying broadcasting user groups based only on static features or set fixed rules in related technologies, and obtain more accurate illegal user group mining results. In addition, on the basis of accurately identifying illegal user groups, it is also helpful to efficiently deal with illegal users, improve the platform's governance capabilities and user experience, and enhance user trust and satisfaction.

[0075] After introducing the basic principles of the present disclosure, various non-limiting embodiments of the present disclosure are described in detail below.

[0076] Application Scenario Overview

[0077] It should be noted that the following application scenarios are only shown to facilitate understanding of the spirit and principle of the present disclosure, and the embodiments of the present disclosure are not limited in this respect. On the contrary, the embodiments of the present disclosure can be applied to any applicable scenario.

[0078] The embodiments of the present disclosure support accurate mining of illegal user groups from the user groups of the creation platform to avoid misidentification of normal creation traffic. For example, in the UGC creation platform, the business connections between resource creation users and resource consumption users can be regularly collected, and then a bipartite graph can be constructed based on their business connections. After that, graph parameters related to the bipartite graph can be collected (graph parameters can include resource creation user characteristics associated with each resource creation user node, resource consumption user characteristics associated with each resource consumption user node, and edge weights of each connecting edge, and the edge weights of each connecting edge are determined based on the traffic contribution of each resource consumption user to the creation resources of resource creation users with whom the resource consumption user has business connections). Thus, illegal user group mining can be performed in combination with the multi-dimensional graph parameters and the above-mentioned bipartite graph to obtain more accurate illegal user group mining results.

[0079] Exemplary Methods

[0080] The exemplary embodiment of the present disclosure first provides a user group mining method. Figure 1 A flowchart of a user group mining method in an embodiment of the present disclosure is shown, which may include the following steps S110 to S130:

[0081] Before step S110, it should be noted that the resource creation users in this disclosure refer to users who create, produce and publish various resources in the UGC platform. These resources are not limited to text, images, audio, video, software code, design templates, teaching materials, etc. Resource creation users can provide valuable content to the platform or other users through their own creativity and professional knowledge, thereby promoting the dissemination and exchange of resources. Therefore, the resource creation users in this disclosure can be video creation users, audio creation users, etc., and can be set according to actual conditions. This disclosure does not make special restrictions on this. Resource creation users are associated with the creative resources they have shared or published.

[0082] Resource consumption users in this disclosure refer to users who use, watch, read or download various resources created and shared by other users on the UGC platform. Resource consumption users are usually recipients and beneficiaries of content. They meet their own needs by accessing and using these resources, such as learning knowledge, entertainment, and obtaining information. Therefore, resource consumption users in this disclosure can be video players, video downloaders, picture browsers, audio listeners, etc., which can be set according to actual conditions, and this disclosure does not make special restrictions on this. For example, assuming that resource creation user A publishes creation resource A, and within a preset period of time (for example: 1 month) after the creation resource A is published, users B, C, and D browse or interact with the above creation resource A, then it can be determined that users B, C, and D have business connections with the creation resources of the above resource creation user A, and users B, C, and D are all resource consumption users of the above creation resource A.

[0083] In step S110, a bipartite graph is constructed according to the business relationship between resource creation users and resource consumption users.

[0084] In this step, the business connections between resource creation users and resource consumption users may be collected, and then a bipartite graph may be constructed based on the business connections between resource creation users and resource consumption users.

[0085] The above-mentioned business connection refers to the business relationship between resource creation users and resource consumption users due to resource browsing or resource interaction (for example: resource access, downloading, likes / favorites, comments, sharing / forwarding).

[0086] A bipartite graph is a special graph structure in which the node set can be divided into two non-overlapping subsets (i.e., resource creation users and resource consumption users in the present disclosure), and each edge in the graph connects the nodes in these two subsets.

[0087] Specifically, the interaction data between all resource creation users and resource consumption users in the UGC platform within a preset period of time (for example, within one month) can be collected. These data may include, but are not limited to: user ID (a unique identifier for each user), resource ID (a unique identifier for each resource), interaction records (recording the resource consumption user's access, download, like, comment, etc. of a resource), and timestamp (the time of each interaction). Furthermore, the collected data can be cleaned and preprocessed (for example, deduplication, missing value processing, format unification, etc.) to ensure the quality and consistency of the data. Afterwards, a graph database or graph algorithm library can be used to construct a bipartite graph, which may include a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes with business connections.

[0088] In step S120, the graph parameters associated with the bipartite graph are determined based on the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge.

[0089] In this step, the graph parameters associated with the bipartite graph can be determined based on the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge.

[0090] Among them, the resource creation user characteristics associated with each of the above-mentioned resource creation users may include the following characteristics: the resource popularity of each resource creation user (also known as the node weight of the resource creation user node), and the resource creation user attributes associated with each resource creation user.

[0091] The resource popularity of each resource creation user can be determined in the following ways:

[0092] refer to Figure 2 , Figure 2 A flowchart showing how to determine the resource popularity of each resource creation user in the embodiment of the present disclosure includes steps S201 and S202:

[0093] In step S201, the average consumption flow of the creation resource of the resource creation user within a preset time period is calculated.

[0094] In this step, the preset time period may be one month, and the duration may be set according to actual conditions, and the present disclosure does not make any special limitation on this.

[0095] Specifically, the consumption traffic can be playback traffic, browsing traffic, download traffic, etc., which can be determined according to actual interactive behavior, and the present disclosure does not make any special limitations on this.

[0096] Therefore, for example, taking the resource creation user as resource creation user a, resource creation user a is associated with creation resources (which can be one or more creation resources), and thus, the total consumption flow of the resource creation user's creation resources within the preset time period can be obtained first. Specifically, all resource consumption users who have business contacts with resource creation user a within one month can be collected (assuming that there are a total of j resource consumption users, then each resource consumption user can be represented as b1, b2, ..., bj), and then, the partial consumption flow of each of the above resource consumption users for the above creation resources within the preset time period can be obtained, and the j partial consumption flows are accumulated to obtain the total consumption flow of all resource consumption users who have business contacts with resource creation user a for the above creation resources within the preset time period.

[0097] Therefore, the above total consumption flow can be expressed as the following formula 1:

[0098] UC a =UP ab1 +UP ab2 +UP ab3 +...+UP abj Formula 1

[0099] Among them, UC a Represents the total consumption traffic of all resource consuming users for one or more created resources of resource creator user a, UP abj Represents the partial consumption traffic of the jth resource consuming user for one or more creative resources of resource creating user a.

[0100] After the total consumption flow is calculated based on the above formula 1, illustratively, the total consumption flow may be divided by the total number of days in the above preset period (eg, 30 days) to obtain the average consumption flow.

[0101] In step S202, the resource heat is determined according to the average consumption flow.

[0102] In this step, the average consumption flow calculated in the above step S201 can be determined as the resource popularity of the above resource creation user a.

[0103] In addition, the resource creation user attributes associated with each resource creation user may include the following attributes: basic attributes (for example: user name, IP address, registration time, last login time, contact information, avatar information, profile information, location information, authentication information, account status, account violation record, login device information, etc.), creation level (can be determined based on creation quality, creation frequency, content influence, etc.), creation quality stratification (can be determined based on creation quality, user feedback effect, etc.), number of creation resources, whether the resource creation user is a popular creation user and other dimensional features, which are used to distinguish between popular resource creation users and ordinary resource creation users, and can be set according to actual conditions. The present disclosure does not make any special limitations on this.

[0104] The resource consumption user characteristics associated with each resource consumption user mentioned above may include the following characteristics: user activity of each resource consumption user (also referred to as node weight of a resource consumption user node), and resource consumption user attributes associated with each resource consumption user.

[0105] The user activity of each resource consumer mentioned above can be determined in the following ways:

[0106] refer to Figure 3 , Figure 3A flow chart showing how to determine the user activity of each resource consuming user in an embodiment of the present disclosure includes steps S301 and S302:

[0107] In step S301, the average consumption flow of the resource consuming user for the historical consumption resources within a preset time period is obtained.

[0108] In this step, the historical consumption resources consumed by the resource consumption user b in the above preset time period (for example, played, clicked, browsed) can be obtained first. The historical consumption resources can be associated with multiple resource creation users (assuming that there are a total of i resource creation users, each resource creation user can be represented as a1, a2, a3, ..., ai), so that the total consumption flow of the resource consumption user for the above historical consumption resources can be obtained. The total consumption flow can be expressed as the following formula 2:

[0109] UP b =UP ba1 +UP ba2 +UP ba3 +...+UP bai Formula 2

[0110] Among them, the above UP b Represents the total consumption flow of resource consumer b for the above historical consumption resources, UP bai Represents the portion of consumption traffic consumed by resource consuming user b for the creation resources of resource creating user ai during the above preset time period.

[0111] After calculating the total consumption flow based on Formula 2, for example, the total consumption flow can be divided by the total number of days in the preset period (for example, 30 days) to obtain the average consumption flow of the resource consumption user for the historical consumption resources in the preset period.

[0112] In step S302, the user activity is determined according to the average consumption flow.

[0113] In this step, the average consumption flow calculated in the above step S301 can be determined as the above user activity.

[0114] In addition, the resource consumption user attributes associated with each resource consumption user may include the following attributes: basic attributes (for example: user name, IP address, registration time, last login time, contact information, avatar information, profile information, location information, authentication information, account status, account violation record, login device information, etc.), the level of the resource consumption user (for example, it can be determined based on the user's registration time, points, frequency of interactive behavior, content contribution, etc.), the behavior type characteristics of the resource consumption user (for example, it can include based on the user's login frequency, online time, consumption behavior, etc.), whether the resource consumption user purchases a membership and other dimensional characteristics, which can be set according to actual conditions, and the present disclosure does not make any special limitations on this.

[0115] When measuring the edge weight of each edge, the present disclosure creatively introduces a calculation method similar to the HHI index (Herfindahl-Hirschman Index (HHI), also known as the Herfindahl Index, which is a comprehensive index for measuring industry concentration. It refers to the sum of the squares of the percentage of total industry revenue or total assets occupied by each market competitor in an industry, and is used to measure changes in market share, that is, the dispersion of manufacturer scale in the market. The Herfindahl Index is one of the better indicators for measuring industry market concentration, and is an indicator widely used by the economics community and government regulatory departments). The total playback traffic of all creative resources under the name of the resource creator user within a preset time period is regarded as the total assets of the industry, and the part of the playback traffic of each resource consuming user playing the resource creator user within the preset time period is regarded as the assets occupied by a competitive entity. Then, the edge weight of each edge is calculated based on the ratio of the latter to the former.

[0116] Specifically, in the present disclosure, the edge weight of each edge is determined according to the traffic contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business connection. Exemplarily, the traffic contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business connection can be determined according to the following method:

[0117] refer to Figure 4 , Figure 4 A flow chart showing how to determine the traffic contribution of each resource consuming user to the resource creating user with whom the user has business contact in the embodiment of the present disclosure includes steps S401 and S402:

[0118] In step S401, the total consumption flow of the creation resources of the resource creation user within a preset time period is obtained, and the partial consumption flow of the creation resources of each resource consumption user within the preset time period is obtained.

[0119] In this step, the total consumption flow UC of the creation resource of the resource creation user within the preset time period can be determined by referring to the relevant description of the above step S201. a , and determine the consumption flow of each resource consuming user for the above-mentioned creative resources within a preset period of time abj , I will not go into details here.

[0120] In step S402, based on the ratio between each partial consumption flow and the total consumption flow, the flow contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business relationship is determined.

[0121] In this step, the traffic contribution of each resource consuming user to the creative resources of the resource creating users with whom the user has business relations can be determined according to the ratio between each part of the consumption traffic and the total consumption traffic.

[0122] Specifically, the traffic contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business connection may be determined based on the following formula 3:

[0123]

[0124] Among them, the above HHI abj It represents the traffic contribution of resource consuming user bj to the resource creating user a with whom he has business relationship.

[0125] It should be noted that in the present disclosure, the sum of the weights of all edges connected to each resource creation user node (which can be expressed as HHI a ) is 1, that is, HHI a =HHI ab1 +HHI ab2 +...+HHI abj =1.

[0126] The present disclosure determines the weight of the resource creation user node, the weight of the resource consumption user node and the weight of each edge through the above scheme, which has at least the following advantages:

[0127] First, the edge weight is the traffic contribution of resource consumption users to resource creation users HHI abj , which facilitates the subsequent removal of edges with small traffic contribution;

[0128] Second, the weight of the resource consumption user node is UP b , which is convenient for eliminating resource-consuming users with small resource consumption (users with small resource consumption do not meet the original intention of broadcasting);

[0129] Third, the weight of the resource creation user node is the total consumption traffic UC of the resource creation user for the preset period of time. a , which is convenient for eliminating creative users with small total consumption traffic (small total consumption traffic does not meet the original intention of broadcasting);

[0130] Fourth, it can effectively distinguish resource consumption users who broadcast resources and resource creation users, because the HHI of resource consumption users who broadcast resources is abj Relatively average and large values, reflecting that the number of accounts for broadcasting is limited, the proportion of broadcasting volume of each account is relatively even, and the traffic contribution to resource creation users is large, while the HHI of resource consumption users who do not broadcast is abj Smaller, and less traffic contribution to resource creators;

[0131] Fifth, it can effectively distinguish between real popular resource creators and non-popular resource creators who broadcast non-popular resources. When the playback volume of the two is similar, the HHI of resource creators who broadcast resource consumers is abj The average value will be much larger than the HHI of the connection between popular resource creator users and resource consumption users. abj The average value is reflected in the fact that the resource consumption users of popular creative users are more dispersed, rather than a few users concentrating on brushing the volume. Therefore, a clear distinction can be made in the mining of illegal gangs.

[0132] In addition, it should be noted that the present disclosure can divide the above-mentioned graph parameters into two parts, namely the first part parameters and the second part parameters. The first part parameters may include: the user activity of each resource consuming user, the resource popularity of each resource creating user and the weight of each of the above-mentioned edges (that is, the traffic contribution of each resource consuming user to the creative resources of the resource creating users with whom he has business relations); the above-mentioned second part parameters may include: the resource consuming user attributes of each resource consuming user and the resource creating user attributes of each resource creating user.

[0133] After the above-mentioned bipartite graph is constructed and its graph parameters are determined, the process may proceed to step S130 to mine illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results.

[0134] In this step, in an optional implementation, the illegal user group mining can be directly performed based on the above-mentioned graph parameters and the above-mentioned bipartite graph through a preset graph mining algorithm to obtain the illegal user group mining results. Exemplarily, the above-mentioned illegal user group can be a resource creation user-resource consumption user group that broadcasts illegally on the UGC platform.

[0135] Among them, the above-mentioned preset graph mining algorithm includes at least one of the following: community discovery algorithm and deep learning algorithm. Exemplarily, the above-mentioned community discovery algorithm may include LPA (Label Propagation Algorithm, a community detection algorithm based on label propagation), Louvain Method (an efficient modularity optimization algorithm), Infomap (using the minimum description length principle in information theory to discover community structure by random walks in the encoding network), ModularityOptimization (finding the best community structure by maximizing an indicator called modularity), etc., which can be set according to actual conditions, and the present disclosure does not make special restrictions on this. The above-mentioned deep learning algorithms may include GraphSAGE (GraphSample and Aggregate, an inductive GNN, which generates feature representations of nodes by sampling and aggregating information from the neighborhood of each node, and is particularly suitable for processing large-scale graph data), GNNs (Graph Neural Networks, neural network models that can operate directly on graph structures, which capture dependencies between nodes through a message passing mechanism), GCNs (Graph Convolutional Networks, which aggregate neighbor information by performing convolution operations in the network to generate feature representations for each node), etc., which can be set according to actual conditions, and the present disclosure does not make any special limitations on this.

[0136] In another optional implementation, given that the above-mentioned graph parameters include the first part parameters and the second part parameters, the constructed bipartite graph can be pruned first according to the first part parameters, and then the illegal user group mining is performed based on the second part parameters and the pruned bipartite graph to obtain the illegal user group mining results.

[0137] In bipartite graphs, pruning refers to simplifying the graph structure by removing certain nodes or edges, thereby improving the efficiency of the algorithm or improving the quality of the results. The main purpose of pruning is to reduce the complexity of the graph, remove unnecessary parts, and make subsequent analysis or processing more efficient and accurate.

[0138] For details, please refer to Figure 5 , Figure 5 A schematic diagram showing another process of obtaining illegal user group mining results in an embodiment of the present disclosure includes steps S501 and S502:

[0139] In step S501, the bipartite graph is pruned based on the first part of parameters to obtain a core bipartite graph.

[0140] In this step, the bipartite graph may be pruned based on the first part of the graph parameters to obtain a pruned core bipartite graph.

[0141] Among them, the above-mentioned first part of parameters includes the user activity of each resource consuming user (i.e., the weight of each resource consuming user node), the resource popularity of each resource creating user (i.e., the weight of each resource creating node), and the traffic contribution of each resource consuming user to the creative resources of the resource creating users with whom he has business relations (i.e., the weight of each edge).

[0142] Specifically, please refer to Figure 6 , Figure 6 A schematic diagram showing how to prune a bipartite graph based on the first part of parameters in an embodiment of the present disclosure to obtain a core bipartite graph includes steps S601 to S605:

[0143] In step S601, the resource heat associated with each resource creation user node is compared with a preset heat threshold, and when the resource heat is not greater than the preset heat threshold, the resource creation user node is pruned.

[0144] In this step, the resource heat associated with each resource creation user node can be compared with the preset heat threshold R1. When the resource heat is not greater than the preset heat threshold R1 (that is, less than or equal to R1, R2 can be set to about one thousand levels, and the specific value of R1 can be set or adjusted according to business needs, and the present disclosure does not make special limitations on this), the current resource creation user node can be removed from the bipartite graph.

[0145] Based on this step, the present disclosure can remove resource creation user nodes with low resource popularity from the bipartite graph, thereby avoiding the misidentification of resource creation users with low resource popularity as illegal users.

[0146] In step S602, when the resource heat is greater than a preset heat threshold, the user activity associated with each resource consuming user node is compared with the preset activity threshold.

[0147] In this step, when the above-mentioned resource heat is greater than the preset heat threshold R1, the current resource creation user node can be retained, and then the user activity associated with each resource consumption user node is compared with the preset activity threshold R2 (that is, less than or equal to R2, R2 can be set to about one hundred, and the specific value of R2 can be set or adjusted according to business needs, and the present disclosure does not make any special limitations on this).

[0148] In step S603, when the user activity is not greater than the preset activity threshold, the resource consuming user node is pruned.

[0149] In this step, when the user activity is not greater than the preset activity threshold R2, the resource consuming user node can be removed from the bipartite graph.

[0150] Based on this step, the present disclosure can remove resource consuming user nodes with low user activity from the bipartite graph, thereby avoiding the misidentification of resource consuming users with low user activity as illegal users.

[0151] In step S604, when the user activity is greater than a preset activity threshold, the traffic contribution associated with each edge is compared with the preset traffic contribution threshold.

[0152] In this step, if the user activity is greater than the preset activity threshold R2, the current resource consumption user node can be retained, and then the traffic contribution associated with each edge is compared with the preset traffic contribution threshold R3 (for example, in the song scenario, the R3 level can be set at 10 -2 level, which can be adjusted as needed in other short video scenes) for comparison.

[0153] In step S605, when the traffic contribution is not greater than a preset traffic contribution threshold, the edge is pruned.

[0154] In this step, when the above-mentioned traffic contribution is not greater than the preset traffic contribution threshold R3, the current edge and the resource consuming user node connected to the edge can be removed from the bipartite graph, and when the traffic contribution is greater than the above-mentioned preset traffic contribution threshold R3, the current edge can be retained.

[0155] The smaller the edge weight is, the smaller the traffic contribution of the resource consuming user to the resource creating user is, and thus, the greater the probability that the resource consuming user is normal traffic. Therefore, based on this step, the present disclosure can effectively eliminate the weight of the edge formed by normal traffic, avoiding accidental damage to normal traffic users in the subsequent identification process.

[0156] refer to Figure 7 , Figure 7 The overall flow chart of how to prune a bipartite graph to obtain a core bipartite graph in the embodiment of the present disclosure includes steps S701 to S710:

[0157] In step S701, start;

[0158] In step S702, the edge weight of the connecting edge is constructed;

[0159] In step S703, the resource heat associated with each resource creation user node is compared with a preset heat threshold to determine whether the resource heat is greater than the preset heat threshold R1;

[0160] In step S704, when the resource heat is not greater than the preset heat threshold R1, the resource creation user node is pruned (the resource creation user node is removed);

[0161] In step S705, when the resource heat is greater than the preset heat threshold R1, the user activity associated with each resource consumption user node is compared with the preset activity threshold R2 to determine whether the user activity is greater than the preset activity threshold R2;

[0162] In step S706, when the user activity is not greater than the preset activity threshold R2, the resource consuming user nodes are pruned (resource consuming user nodes are removed);

[0163] In step S707, when the user activity is greater than the preset activity threshold, the traffic contribution associated with each edge is compared with the preset traffic contribution threshold to determine whether the traffic contribution is greater than the preset traffic contribution threshold R3;

[0164] In step S708, when the traffic contribution is not greater than the preset traffic contribution threshold R3, the edge is pruned (the edge is removed);

[0165] In step S709, when the traffic contribution is greater than a preset traffic contribution threshold, the edge is retained;

[0166] In step S710, end.

[0167] Next reference Figure 5 In step S502, illegal user group mining is performed based on the second part of parameters and the core bipartite graph to obtain illegal user group mining results.

[0168] In this step, the illegal user group mining can be directly performed based on the second part of the parameters (the resource creation user attributes and resource consumption user attributes) and the core bipartite graph through a preset graph mining algorithm to obtain the illegal user group mining results. Exemplarily, the illegal user group mining results may include the resource creator-resource consumer set of the swiping, and may also include relevant explanations or reasons for being the illegal user group.

[0169] refer to Figure 8 , Figure 8 The overall flow chart of the user group mining method in the embodiment of the present disclosure is shown, including steps S801 to S804:

[0170] In step S801, a connection relationship is established: a connection relationship between a resource creation user and a resource consumption user may be established first. Specifically, the weight of each resource creation user node, the weight of each resource consumption user node, and the weight of each edge may be calculated (determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations);

[0171] In step S802, pruning: pruning the bipartite graph according to the weight of each resource creation user node, the weight of each resource consumption user node and the weight of each edge;

[0172] In step S803, features are added: resource creation user attributes associated with each resource creation user, and resource consumption user attributes associated with each resource consumption user;

[0173] In step S804, graph mining: using a community discovery algorithm and / or a deep learning algorithm to determine mining results of illegal user groups.

[0174] The present disclosure provides a method for group mining in UGC playback scenarios that can effectively exclude natural traffic such as fan check-ins and normal resource consumption, and identify brush traffic in response to the ever-changing brushing techniques. This method constructs the HHI edge weights between resource creation users and resource consumption users to measure the contribution of resource consumption users to the traffic of resource creation users' works. At the same time, it introduces the attribute characteristics of resource creation users and resource consumption users to effectively filter natural traffic and fan check-ins for popular resource creation users, and effectively mine brush traffic. In UGC brushing scenarios with various cheating techniques, this method has strong robustness and compatibility.

[0175] Based on the above technical solutions, the present disclosure has at least the following technical effects:

[0176] First, unlike the previous solution that directly uses static features such as the player's IP and device to compose the graph, this disclosure builds a graph association relationship based on the traffic contribution (HHI index) between resource playback users and resource creation users in the case of complex playback behaviors and diverse traffic patterns (natural traffic, fan check-ins, black production, etc.) in UGC scenes, which can effectively eliminate the impact of natural traffic;

[0177] Second, by introducing the activity of resource consumption users, it is possible to effectively distinguish between zombie users who broadcast and ordinary resource consumption users; by introducing the resource popularity of resource creation users, it is possible to effectively distinguish between popular creation users and ordinary creation users;

[0178] Third, by combining the constructed bipartite graph with the graph parameters (the user activity of each resource consuming user, the resource popularity of each resource creating user and the weight of each of the above-mentioned edges (i.e., the traffic contribution of each resource consuming user to the creative resources of the resource creating users with whom he has business connections), the resource consuming user attributes of each resource consuming user and the resource creating user attributes of each resource creating user), and mining the illegal user groups through community discovery algorithms and / or deep learning algorithms, the accuracy of identifying the UGC scenario brushing groups can be improved.

[0179] Exemplary Devices

[0180] After introducing the user group mining method according to the exemplary embodiment of the present disclosure, Fig. 9 The user group mining device 900 according to the exemplary embodiment of the present disclosure is described. The user group mining device 900 includes:

[0181] A graph construction module 910 is used to construct a bipartite graph according to the business relationship between resource creation users and resource consumption users; the resource creation users are associated with creation resources, and the bipartite graph includes a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes having business relationships;

[0182] A graph parameter determination module 920 is used to determine the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge; the edge weight of each connecting edge is determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations;

[0183] The user group mining module 930 is used to mine illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results.

[0184] In an optional embodiment, the graph parameters associated with the bipartite graph include first part parameters and second part parameters;

[0185] The user group mining module 930 mines illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results, including:

[0186] Pruning the bipartite graph based on the first part of parameters to obtain a core bipartite graph;

[0187] Based on the second part of parameters and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

[0188] In an optional implementation, the resource consumption user characteristics include the user activity of each resource consumption user, the resource creation user characteristics include the resource popularity of each resource creation user, and the first part of parameters includes the user activity, the resource popularity and the traffic contribution;

[0189] The user group mining module 930 performs pruning processing on the bipartite graph based on the first part of parameters, including:

[0190] Comparing the resource heat associated with each resource creation user node with a preset heat threshold, and performing pruning processing on the resource creation user node when the resource heat is not greater than the preset heat threshold;

[0191] When the resource heat is greater than the preset heat threshold, comparing the user activity associated with each of the resource consuming user nodes with the preset activity threshold;

[0192] When the user activity is not greater than the preset activity threshold, pruning the resource consuming user node;

[0193] When the user activity is greater than the preset activity threshold, comparing the traffic contribution associated with each edge with the preset traffic contribution threshold;

[0194] When the traffic contribution is not greater than a preset traffic contribution threshold, pruning is performed on the edge.

[0195] In an optional implementation, the traffic contribution of each resource consuming user to the creative resources of the resource creating user with whom the user has business connection is calculated in the following way:

[0196] Obtaining the total consumption flow of the creation resource of the resource creation user within a preset time period, and obtaining the partial consumption flow of each resource consumption user for the creation resource within the preset time period;

[0197] According to the ratio between each of the partial consumption flows and the total consumption flow, the flow contribution of each resource consuming user to the creation resources of the resource creation users with whom the user has business relations is determined.

[0198] In an optional implementation, the resource heat is calculated in the following way:

[0199] Calculate the average consumption flow of the creation resource of the resource creation user within a preset time period;

[0200] The resource heat is determined according to the average consumption flow.

[0201] In an optional implementation manner, the user activity is calculated in the following manner:

[0202] Obtaining the average consumption flow of the resource consuming user for historical consumption resources within a preset time period;

[0203] The user activity is determined according to the average consumption flow.

[0204] In an optional implementation, the resource consumption user characteristics further include resource consumption user attributes, the resource creation user characteristics further include resource creation user attributes, and the second part of parameters includes the resource consumption user attributes and the resource creation user attributes;

[0205] The user group mining module 930 mines illegal user groups based on the second part of parameters and the core bipartite graph to obtain illegal user group mining results, including:

[0206] Based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

[0207] In an optional implementation, the user group mining module 930 performs illegal user group mining based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph to obtain illegal user group mining results, including:

[0208] Performing user group mining based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph through a preset graph mining algorithm to obtain illegal user group mining results;

[0209] Among them, the preset graph mining algorithm includes at least one of the following: a community discovery algorithm and a deep learning algorithm.

[0210] In addition, other specific details of the embodiments of the present disclosure have been described in detail in the inventive implementation of the above method and will not be repeated here.

[0211] Exemplary Storage Media

[0212] The storage medium according to the exemplary embodiment of the present disclosure is described below.

[0213] In this exemplary embodiment, the above method can be implemented by a program product, such as a portable compact disk read-only memory (CD-ROM) and including program code, and can be run on a device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, apparatus, or device.

[0214] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0215] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0216] The program code contained on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RE, etc., or any suitable combination of the foregoing.

[0217] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user computing device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (FAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0218] Exemplary Electronic Devices

[0219] refer to Fig.10 An electronic device according to an exemplary embodiment of the present disclosure is described.

[0220] Fig.10 The electronic device 1000 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0221] like Fig.10 As shown, the electronic device 1000 is in the form of a general computing device. The components of the electronic device 1000 may include but are not limited to: at least one processor 1010, at least one memory 1020, a bus 1030 connecting different system components (including the memory 1020 and the processor 1010), and a display unit 1040.

[0222] The memory stores program codes, which can be executed by the processor 1010, so that the processor 1010 executes the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification. For example, the processor 1010 can execute the following steps: Figure 1 The method steps shown, etc.

[0223] The memory 1020 may include a volatile memory, such as a random access memory (RAM) 1021 and / or a cache memory 1022 , and may further include a read-only memory (ROM) 1023 .

[0224] Memory 1020 may also include a program / utility 1024 having a set (at least one) of program modules 1025, such program modules 1025 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0225] The bus 1030 may include a data bus, an address bus, and a control bus.

[0226] The electronic device 1000 may also communicate with one or more external devices 1100 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), and such communication may be performed through an input / output (I / O) interface 1050. The electronic device 1000 also includes a display unit 1040, which is connected to the input / output (I / O) interface 1050 for display. In addition, the electronic device 1000 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter 1060. As shown, the network adapter 1060 communicates with other modules of the electronic device 1000 through a bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1000, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0227] It should be noted that, although several modules or submodules of the device are mentioned in the above detailed description, such division is merely exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.

[0228] In addition, although the operations of the disclosed method are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0229] Although the spirit and principle of the present disclosure have been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the disclosed specific embodiments, and the division of various aspects does not mean that the features in these aspects cannot be combined to benefit, and such division is only for the convenience of expression. The present disclosure is intended to cover various modifications and equivalent arrangements included in the spirit and scope of the attached claims.

Claims

1. A user group mining method, characterized in that: include: Construct a bipartite graph based on the business relationship between resource creation users and resource consumption users; The resource creation user is associated with a creation resource, and the bipartite graph includes a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes having business connections; Determine the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge; the edge weight of each connecting edge is determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations; Based on the graph parameters and the bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

2. The method according to claim 1, characterized in that The graph parameters associated with the bipartite graph include first part parameters and second part parameters; The mining of illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results includes: Pruning the bipartite graph based on the first part of parameters to obtain a core bipartite graph; Based on the second part of parameters and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

3. The method according to claim 2, characterized in that The resource consumption user characteristics include the user activity of each resource consumption user, the resource creation user characteristics include the resource popularity of each resource creation user, and the first part of parameters includes the user activity, the resource popularity and the traffic contribution; The pruning of the bipartite graph based on the first part of parameters includes: Comparing the resource heat associated with each of the resource creation user nodes with a preset heat threshold, and performing pruning processing on the resource creation user node when the resource heat is not greater than the preset heat threshold; When the resource heat is greater than the preset heat threshold, comparing the user activity associated with each of the resource consuming user nodes with the preset activity threshold; When the user activity is not greater than the preset activity threshold, pruning the resource consuming user node; When the user activity is greater than the preset activity threshold, comparing the traffic contribution associated with each edge with the preset traffic contribution threshold; When the traffic contribution is not greater than a preset traffic contribution threshold, pruning is performed on the edge.

4. The method according to claim 3, characterized in that The traffic contribution of each resource consuming user to the resource creating user with whom he has business relationship is calculated as follows: Obtaining the total consumption flow of the creation resource of the resource creation user within a preset time period, and obtaining the partial consumption flow of each resource consumption user for the creation resource within the preset time period; According to the ratio between each of the partial consumption flows and the total consumption flow, the flow contribution of each resource consuming user to the creation resources of the resource creation users with whom the user has business relations is determined.

5. The method according to claim 3, characterized in that: The resource heat is calculated in the following way: Calculate the average consumption flow of the creation resource of the resource creation user within a preset time period; The resource heat is determined according to the average consumption flow.

6. The method according to claim 3, characterized in that The user activity is calculated in the following way: Obtaining the average consumption flow of the resource consuming user for historical consumption resources within a preset time period; The user activity is determined according to the average consumption flow.

7. The method according to claim 2, characterized in that The resource consumption user characteristics also include resource consumption user attributes, the resource creation user characteristics also include resource creation user attributes, and the second part of parameters includes the resource consumption user attributes and the resource creation user attributes; The mining of illegal user groups based on the second part of parameters and the core bipartite graph to obtain illegal user group mining results includes: Based on the resource consumption user attributes, the resource creation user attributes and the core bipartite graph, illegal user group mining is performed to obtain illegal user group mining results.

8. A user group mining device, characterized in that: include: A graph construction module, used to construct a bipartite graph based on the business relationship between resource creation users and resource consumption users; The resource creation user is associated with a creation resource, and the bipartite graph includes a resource creation user node for representing each resource creation user, a resource consumption user node for representing each resource consumption user, and edges between resource creation user nodes and resource consumption user nodes having business connections; A graph parameter determination module, for determining the graph parameters associated with the bipartite graph according to the resource creation user characteristics associated with each resource creation user node, the resource consumption user characteristics associated with each resource consumption user node, and the edge weight of each connecting edge; the edge weight of each connecting edge is determined according to the traffic contribution of each resource consumption user to the creation resources of the resource creation user with whom the resource consumption user has business relations; The user group mining module is used to mine illegal user groups based on the graph parameters and the bipartite graph to obtain illegal user group mining results.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to perform the method according to any one of claims 1 to 7 by executing the executable instructions.

Images