WO2021217933A1 - Community division method and apparatus for homogeneous network, and computer device and storage medium - Google Patents

Abstract

The embodiments of the present application belong to the field of big data, and relate to a community division method and apparatus for a homogeneous network, and a computer device and a storage medium. The method comprises: receiving financial transaction information; taking a transaction user in the financial transaction information as a node, and constructing a homogeneous network according to a transaction relationship and transaction data; initializing a community tag of each node to obtain a community tag set; performing calculation by means of taking the transaction data as a weight, so as to obtain a node influence value of each node and a relative influence value of each node to an adjacent node; iteratively updating the community tag set of each node according to the node influence value and the relative influence value; and when a community tag in the community tag set of each node no longer changes, dividing the homogeneous network into at least one community node network according to the community tag. By means of the present application, the accuracy of community division is improved. In addition, the present application further relates to blockchain technology, and the financial transaction information can be stored in a blockchain.

Description

Homogeneous network community division method, device, computer equipment and storage medium

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on April 29, 2020, the application number is 202010356641.6, and the invention title is "Methods, Apparatus, Computer Equipment, and Storage Media for Community Division of Homogeneous Networks". The entire content is incorporated into this application by reference.

Technical field

This application relates to the field of big data technology, and in particular to a method, device, computer equipment, and storage medium for dividing a homogeneous network into a community.

Background technique

Homogeneous network refers to a network formed by the interconnection of entities of the same category, which can actually be a network composed of people or computers and computers. People or computers in a homogeneous network can be regarded as nodes, and the actual homogeneous network may have hundreds of millions of nodes. Some nodes in a homogeneous network may form a community node network due to close relationships. The community node network has important practical value, such as promotion based on core users in the community node network, identification of fraudulent gangs, etc., so it is often necessary to obtain information from A network of community nodes is divided into a homogeneous network. The community division of homogeneous networks involves data mining in the field of big data, which usually requires the processing of massive amounts of data.

Existing technologies for dividing homogeneous networks usually use structured analysis methods. Based on existing experience, analysts screen and compare complex and cumbersome transaction forms, which consumes a lot of labor costs. The inventor realizes that the analysis method based on structured data cannot mine the correlation between multiple nodes, loses a lot of connection information, and does not consider the difference in importance between nodes, and cannot accurately identify important nodes. In addition, the existing analysis methods do not overlap the division of nodes, that is, the nodes belong to a single community node network. However, in practice, for example, fraudulent gangs, including multiple roles such as intermediaries, initiators, and destinations, must have nodes that exist in multiple community node networks at the same time, and existing analysis methods cannot determine whether a node is a different community. The overlapping part of the group node network results in a lower accuracy of the homogenous network community division.

Summary of the invention

The purpose of the embodiments of the present application is to propose a method, device, computer equipment, and storage medium for community division in a homogeneous network, so as to solve the problem of low accuracy of community division in a homogeneous network.

A method for community division of a homogeneous network, the method includes:

Receive financial transaction information;

Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.

In order to solve the foregoing technical problems, an embodiment of the present application also provides a community division device for a homogeneous network, including:

Information receiving module for receiving financial transaction information;

The network construction module is configured to use the transaction users in the financial transaction information as nodes, and construct a homogeneous network based on the transaction relationship and transaction data between the transaction users;

The label initialization module is used to initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

A node calculation module, configured to calculate each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

The label update module is configured to iteratively update the community label set of each node according to the node influence value and the relative influence value; there is at least one community label in the community label set after the iterative update;

The network division module is used to divide the homogenous network into at least the same network according to the community label after the iterative update of each node when the community label in the community label set of each node no longer changes. A network of community nodes.

In order to solve the above technical problems, an embodiment of the present application also provides a computer device, including a memory and a processor, and computer-readable instructions stored in the memory and capable of running on the processor, and the processor executes The computer-readable instructions implement the following steps:

Receive financial transaction information;

Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.

In order to solve the above technical problems, embodiments of the present application also provide a computer-readable storage medium, the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions implement the following steps when executed by a processor:

Receive financial transaction information;

Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.

Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: after receiving financial transaction information, the transaction user in the financial transaction information is used as a node, a homogeneous network is constructed based on the transaction relationship and transaction data between the transaction users, and initialized The community label of each node in the homogeneous network is obtained, and the community label set of each node in the homogeneous network is obtained; when calculating the node influence value of each node and the relative influence value of each node on neighboring nodes, the transaction data is used as the weight , Taking into account the importance of the node, the accuracy of the calculation result is improved; according to the node influence value and relative influence value, the community label set of each node is iteratively updated, and there is at least one community in the iteratively updated community label set Label, the community label identifies the degree of belonging of the node to a certain community; when the community label in the community label set of each node no longer changes in the iterative update, the community label after iterative update according to each node will be the same The qualitative network is divided into at least one community node network, and nodes with more than one community label are divided into more than one community node network, which ensures the accuracy of community division of the homogeneous network.

Description of the drawings

The following will briefly introduce the drawings used in the description of the embodiments of the present application. Obviously, the drawings in the following description are some embodiments of the present application.

Figure 1 is an exemplary system architecture diagram to which the present application can be applied;

Fig. 2 is a flowchart of an embodiment of a method for community division of a homogeneous network according to the present application;

Figure 3 is a schematic diagram of a homogeneous network in an embodiment;

FIG. 4 is a flowchart of a specific implementation of step 204 in FIG. 2;

FIG. 5 is a flowchart of another specific implementation manner of step 204 in FIG. 2;

FIG. 6 is a flowchart of a specific implementation of step 205 in FIG. 2;

FIG. 7 is a flowchart of a specific implementation of step 206 in FIG. 2;

FIG. 8 is a schematic structural diagram of an embodiment of a device for dividing a community in a homogeneous network according to the present application;

Fig. 9 is a schematic structural diagram of an embodiment of a computer device according to the present application.

Detailed ways

The terminology used in the specification of the application herein is only for the purpose of describing specific embodiments, and is not intended to limit the application.

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

As shown in FIG. 1, the system architecture 100 may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, and so on. The user can use the terminal devices 101, 102, and 103 to interact with the server 105 through the network 104 to receive or send messages and so on. Various communication client applications can be installed on the terminal devices 101, 102, 103. The terminal devices 101, 102, 103 may be various electronic devices with display screens and support for web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III, dynamic Video experts compress standard audio layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, dynamic image experts compress standard audio layer 4) players, laptop portable computers and desktop computers, etc. The server 105 may be a server that provides various services.

It should be noted that the community division method of the homogeneous network provided in the embodiments of the present application is generally executed by the server, and accordingly, the community division device of the homogeneous network is generally set in the server. It should be understood that the numbers of terminal devices, networks, and servers in FIG. 1 are merely illustrative. There can be any number of terminal devices, networks, and servers according to implementation needs. Continuing to refer to FIG. 2, there is shown a flowchart of an embodiment of a method for community division of a homogenous network according to the present application. The community division method of the homogeneous network includes the following steps:

Step 201: Receive financial transaction information.

Among them, financial transaction information may be information that records financial transactions, including transaction users (user names or user accounts, etc.), transaction relationships between user accounts, and transaction amounts. Financial transaction information may also include transaction time and other information; transactions; The relationship can be to record whether a financial transaction has occurred between transaction users, and the direction of transfer when a financial transaction occurs. Financial transaction information is an important data for macroeconomic research and analysis.

In one embodiment, financial transaction information is stored in the blockchain. In order to ensure the privacy and security of the above-mentioned financial transaction information, the financial transaction information can be stored in a node of a blockchain, and the financial transaction information can be shared between different platforms through the blockchain. Blockchain is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

Specifically, the electronic device (such as the server shown in FIG. 1) on which the community division method of the homogeneous network runs can receive the financial transaction information uploaded by each terminal through various wired connection methods or wireless connection methods.

Various applications or pages that can perform financial transactions are running in the terminal. The user manipulates the terminal to perform financial transactions to generate financial transaction information, and the terminal uploads the financial transaction information to the server.

In one embodiment, the financial transaction information may be stored in a database, and the server obtains the financial transaction information from the database. For example, the server can obtain financial transaction information from the server of a financial transaction platform such as a bank.

In step 202, a transaction user in the financial transaction information is used as a node, and a homogenous network is constructed based on the transaction relationship and transaction data between the transaction users.

Specifically, the server extracts the transaction relationship and transaction data between the transaction user and the transaction user from the financial transaction information. The server regards the transaction users as nodes, and determines the relationship between the nodes corresponding to the transaction users according to the transaction relationship and transaction data between the transaction users, and builds a homogeneous network. If there is a transaction relationship between the transaction users corresponding to the two nodes, the two nodes can be directly connected, and the two nodes are adjacent to each other.

In one embodiment, the step of using transaction users in financial transaction information as nodes and constructing a homogeneous network based on transaction relationships and transaction data between transaction users includes: identifying transaction users in financial transaction information and transactions between transaction users Relationship and transaction data; use the identified transaction users as the nodes in the homogenous network to be constructed; connect the nodes with transaction relationships through the node connection line, and mark the transaction data corresponding to the transaction relationship on the node connection line to obtain the same Quality network.

Specifically, the server recognizes user names in financial transaction information and transaction data between users through a semantic recognition model, and obtains user names and transaction data between users in text format (TXT). The semantic recognition model can use existing Model, such as the BERT model. When constructing a homogeneous network, the user name expressed in text format is loaded into the homogeneous network to be constructed as a node. When there is a transaction relationship between nodes, the two nodes are connected by a node connection line, and the corresponding transaction data is marked on the connection line in a text format, thereby completing the construction of a homogeneous network.

In this embodiment, the identified transaction users are used as nodes, nodes with transaction relationships are connected through node connection lines, and the corresponding transaction data is marked on the node connection lines, ensuring the accuracy of constructing a homogeneous network.

In one embodiment, the financial transaction information is assigned an information identifier, corresponding to a certain range of transaction users. The server reads the information identifier after receiving the financial transaction information, and queries whether the historical financial transaction data carrying the information identifier has been processed. If processing has been performed, extract the stored homogeneous network corresponding to the information identifier from the database, and update the homogeneous network according to the latest obtained financial transaction information.

The homogenous network constructed by the server can be as shown in FIG. 3, and this application uses the specific embodiment in FIG. 3 to explain the technical solution to be disclosed. Specifically, referring to Figure 3, where nodes A, B, C, and D are the four identified transaction users, respectively; the node connection line adopts a directed line segment to identify the transaction relationship between the nodes, and the transaction relationship is represented by numbers. For example, if it is recognized that node A transfers 1500 yuan to node B, a directed line segment from node A to node B is constructed as a node connection line, and the transaction data 1500 is marked on the node connection line.

Step 203: Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network.

Among them, the community label can identify the degree of belonging of the node to a certain community; the community label set can be a set of node community labels.

Specifically, after the homogenous network is constructed, it needs to be initialized. The server assigns different community numbers to the nodes in the homogeneous network, and uniformly sets the attribution degree of the node to the community corresponding to the assigned community number as a default value. The community number corresponds to a community, and a community is a collection of nodes with a certain relationship. A community number and the attribution of the node to the community number constitute a community label of the node, and all the community labels of the node constitute the community label set of the node.

In one embodiment, the step of initializing the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network includes: configuring a community number to each node in the homogeneous network; initializing each node in the homogeneous network; The degree of attribution of the node to the configured community number; based on the community number of each node and the degree of attribution corresponding to the community number, the community label of each node is generated; the community of each node is constructed according to the community label of each node Set of labels.

Specifically, the server allocates a community number to each node in the homogeneous network, and the community number of each node is different from each other. The server initializes the attribution degree of each node to the configured community number and sets it to 1 uniformly. In the later iterative update, the change interval of the attribution degree is [0,1]. The server combines the community number of the node and the attribution degree corresponding to the community number to obtain a community label of the node. The node community label set is the node’s community label set. At the time of initialization, there is only one community label in the community label set of the node.

For example, referring to FIG. 3, during initialization, the server assigns community numbers a, b, c, and d to nodes A, B, C, and D in sequence, and the node's attribution degree to the community number is B=1. Then the community label of node A is l(a,1), and the community label set of node A is {l(a,1)}. Similarly, the community label sets of nodes B, C, and D are: {l(b,1)}, {l(c,1)} and {l(d,1)}.

In this embodiment, the community number and corresponding attribution of each node are initialized, and a community label and a community label set are generated, which ensures the normal progress of subsequent community division.

Step 204: Calculate each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on adjacent nodes.

Among them, the node influence value may be the evaluation value of the influence of the node in the homogeneous network; the relative influence value may be the evaluation value of the influence of the node on an adjacent node in the homogeneous network.

Specifically, nodes exert influence in a homogeneous network through transaction data. The server uses the transaction data as the weight and calculates according to the preset node calculation method to obtain the node influence value of each node in the homogeneous network. Then according to the calculated node influence value, calculate the relative influence value of each node to its neighboring nodes.

Step 205: According to the node influence value and the relative influence value, the community label set of each node is iteratively updated; there is at least one community label in the community label set after the iterative update.

Specifically, the server sorts the nodes in ascending order according to the node influence value, and uses the sort order as the update order of the community tag set of each node. The server updates the community label set of the node based on the community label set of the neighboring nodes of the node.

During the iterative update, the original community label of the node itself does not affect itself. The server obtains the relative influence value of the node's neighboring node on the node and the community label of the neighboring node, and calculates the attribution degree of the node to the community number in the community label of the neighboring node. The server screens from the calculated attribution degrees according to the values, combines at least one attribution degree and its corresponding community number into a new community label of the node, and obtains an iteratively updated community label set. There is at least one community label in the community label set after iterative update, and the community numbers in the community labels are different from each other.

Referring to Fig. 3, assuming that the order of the iterative update of nodes is: A, B, C, D, the server first iteratively updates the community tag set α of node A to obtain α'. Next, when the community label set β of node B is updated, the community label set of node A updates β based on α′.

Step 206: When the community label in the community label set of each node no longer changes during the iterative update, divide the homogeneous network into at least one community node network according to the community label after the iterative update of each node.

Among them, the community node network can be a network composed of nodes that have certain close connections in financial transactions, which is a sub-network of a homogeneous network.

Specifically, after the server completes a round of iterative update for each node in the homogeneous network, it compares whether the current community tag set of each node is consistent with the community tag set after the previous iteration; if it is inconsistent, continue to check The community tag set of each node is updated iteratively; if they are consistent, the community tag set of each node has not changed, then the iteration is stopped, and the community number of the community tag in the current community tag set of each node will be the same The nodes of the community number are included in the same community node network. The server divides at least one community node network from the homogeneous network.

When there is more than one community label in the community label of a node, the node is divided into more than one community node network and becomes an overlapping node.

The community tag set has not changed in the iterative update, including: the number of community tags has not changed, the community number has not changed, and the degree of attribution corresponding to the community number has not changed.

In this embodiment, after receiving the financial transaction information, the transaction users in the financial transaction information are used as nodes, a homogenous network is constructed based on the transaction relationship and transaction data between the transaction users, and the community tags of each node in the homogenous network are initialized to obtain The community label set of each node in the homogeneous network; when calculating the node influence value of each node and the relative influence value of each node on neighboring nodes, the transaction data is used as the weight, taking into account the importance of the node, and improving the calculation result According to the node’s influence value and relative influence value, the community label set of each node is iteratively updated. There is at least one community label in the community label set after iterative update. The community label identifies the node’s influence on a certain community The belonging degree of the group; when the community label in the community label set of each node no longer changes in the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node. Nodes with more than one community label are divided into more than one community node network, which ensures the accuracy of community division for homogeneous networks.

Further, as shown in FIG. 4, in the foregoing step 204, calculating the node influence value may include:

In step 2041, the transaction data is used as the weight, and the transaction weight value of each node is calculated according to the transaction relationship.

Among them, the transaction weight value can be a measure of the influence of the transaction data related to the node in the homogeneous network.

Specifically, referring to Fig. 3, node A has three node connection lines (AB, AC, AD), and the transaction data on the three node connection lines are added to obtain the node transaction amount k _{A of node A.}

Substituting the data in Figure 3 can be obtained: k _A =1000+1500+10000=12500;

Node A has 3 adjacent triangles (ABC, ADC, ABD). The sides of the triangle are node connection lines. The average transaction data on the three node connection lines can be used to obtain the triangle weight of each adjacent triangle, and then the three triangle weights Add up to get the weight e _{A of node A.}

For the adjacent triangle ABC of node A:

e _ABC ＝(1500+10000+20000)/3=10500

The same principle can be obtained: e _ADC = 13666, e _ABD = 4166

e _A ＝10500+13666+4166=28332

In the same way, we can get: e _B =39666, e _C =46666, and e _D =40332.

For each node in the homogeneous network, the transaction amount node k _m e _m and weights are added to obtain a weight value transactions node (k _m + e _m).

Step 2042: For each node in the homogeneous network, filter the maximum transaction weight value and the minimum transaction weight value from the transaction weight values corresponding to the neighboring nodes of the node.

Specifically, the server selects the transaction weight value with the largest value and the transaction weight value with the smallest value from the transaction weight values of the node and its neighboring nodes, and uses the transaction weight value with the largest value as the largest transaction weight value.

Use the smallest transaction weight value as the smallest transaction weight value

For Figure 3, when YIDA calculates node A, among node A and its neighboring nodes, the maximum transaction weight value comes from node C, and the minimum transaction weight value comes from node A, where

Step 2043: Calculate the node influence value of the node in the homogeneous network according to the transaction weight value, the maximum transaction weight value and the minimum transaction weight value of the node.

Specifically, the node influence value of node A can be calculated according to the following formula (1):

Among them, e _A is the weight _{of node A, k A} is the node transaction amount of node A,

Is the maximum transaction weight value in node A and its neighboring nodes,

Is the minimum transaction weight value in node A and its neighboring nodes.

Substitute specific data to get:

According to this, the node influence value of node A can be obtained as 0.5. The server calculates the node influence value of the remaining nodes in the same way.

In this embodiment, when performing node analysis and calculating the node influence value, the transaction data is substituted into the calculation as the weight, which takes into account the differences in financial transactions between the users corresponding to the node, and improves the accuracy of the calculated node influence value. .

Further, as shown in FIG. 5, in the foregoing step 204, calculating the relative influence value of each node on the neighboring node may include:

Step 2044: Calculate the similarity between each pair of adjacent nodes according to the transaction relationship and transaction data.

Wherein, the similarity may be an evaluation value of the similarity of two nodes that are adjacent to each other.

Specifically, when calculating the similarity between each pair of adjacent nodes, it is necessary to calculate the path weight between the two adjacent nodes according to the transaction relationship and transaction data. The calculation of the path weight needs to consider the first-degree relationship and the second-degree relationship of adjacent nodes.

Referring to Figure 3, node A and node C are adjacent nodes to each other. When node A and node C are directly connected by a node connection line, they belong to the first-degree relationship; when node A first passes through node D (or node B), and then arrives At node C, it belongs to the two-degree relationship between the two.

For node A and node C in Figure 3, there is AC when the relationship is first degree, and the path weight of the first degree is: for a node connection line between the first network node A and the second network node C, that is, AC, then we get The weight of the first-degree path ∑w _AC1 ＝10000;

When node A and node C are in the second-degree relationship, there are ADC and ABC, and the second-degree path weight is obtained at this time

Then the path weight of node A relative to node C is s(A, C)=10000+13125=23125.

Let ∑ _x∈Ng(A) s(A,x) represent the sum of s of all adjacent nodes of node A relative to A, then:

∑ _x∈Ng(A) s(A,x)=s(A,D)+s(A,C)+s(A,B)=15750+23125+13625=52500,

In the same way, we can get: ∑ _y∈Ng(C) s(C,y)=s(C,D)+s(C,A)+s(C,B)=42125+23125+34750=100,000.

The server calculates the similarity Sim between node A and node C according to the following formula (2):

Substitute the above data into:

0.32 is the similarity between node A and node C, and the server calculates the similarity between adjacent nodes in the same way. Node C has two other adjacent nodes B and D. The same principle can be obtained: Sim(B,C)=0.39, Sim(D,C)=0.45.

Step 2045: For each node in the homogeneous network, determine the maximum similarity of adjacent nodes from the similarity corresponding to the adjacent nodes of the node.

Specifically, there is similarity between node C and node A, there is similarity between node C and node B, and there is similarity between node C and node D. When node A is used as the main body to calculate the relative influence value of node A on adjacent node C, from the three similarities corresponding to node C, according to the numerical value, the largest similarity is selected as the maximum similarity of node C

It is easy to find that the similarity between node C and node D is 0.45, which is the maximum similarity of node C.

Step 2046: Based on the node influence value of the node, the similarity between the node and the adjacent node, and the maximum similarity of the adjacent node, the relative influence value of the node on the adjacent node is calculated.

Specifically, referring to Fig. 3, the relative influence value NNI _A (C) of node A on node C is calculated according to the following formula (3):

Among them, NI _A is the node influence value of node A, Sim(A,C) is the similarity between node A and node C, h is the adjacent node of node C,

Is to take the maximum similarity between node C and its neighboring nodes, where node D and node C have the maximum similarity of 0.45. Substituting the data can get:

In this embodiment, the similarity between adjacent nodes is calculated on the basis of the node influence value, and the relative influence value of the node on the adjacent node is calculated on the basis of the similarity and the node influence value. When calculating the node influence value, the transaction The data is used as the weight to ensure the accuracy of the node's influence value, thereby ensuring the accuracy of the final calculation of the relative influence value.

Further, as shown in FIG. 6, the foregoing step 205 may include:

Step 2051: Sort the nodes in ascending order according to the node influence value to obtain the update order sequence.

Specifically, before iteratively updating the community tag set of the node, the update order is determined first. The server sorts the nodes in ascending order according to the numerical value of the node's influence value, and uses the sorted ordered sequence as the update sequence sequence. In the update sequence sequence, the arrangement order of the nodes determines the update order of the node community tag set. The node with the largest node influence value is finally updated iteratively to ensure that its community label can be spread to the greatest extent and affect other nodes.

Step 2052: For each node in the update sequence sequence, obtain the relative influence value of each adjacent node of the node on the node and the first label set of the node in turn; the first label set is composed of the community label set of each adjacent node with the largest attribution Degree of community label composition.

Specifically, when the server performs the iterative update according to the update sequence sequence, for the node that currently needs to be updated, it queries its neighboring nodes. From the community label set of each neighboring node, the server separately searches for the community label with the greatest degree of attribution, and combines the searched community labels to obtain the node’s first label set; the first label set is used to The community label set is updated. The server also needs to obtain the relative influence value of each neighboring node of the node on the node.

Each community label set has a main label, and the main label is the community label with the largest attribution in the community label set, that is, the first label set is composed of the main labels of each neighboring node.

For node u∈{A,B,C,D}, when updating the community label set of node u, first obtain the main label set L _Ng of the neighboring nodes of node u, that is, the first label set:

L _Ng = {l(c ₁ ,B ₁ ),l(c ₂ ,B ₂ ),...l(c _v ,B _v )},v∈Ng(u)

Among them, l(c _v , B _v ) represents the main label of the adjacent node v, and B _v represents the degree of membership of the node v to the community number c _{v after the last iteration.}

Step 2053: Based on the relative influence value, calculate the attribution degree of the node to the community number in each community label in the first label set to obtain the second label set.

Specifically, the attribution degree of the node to the community number in each community label in the first label set is calculated by formula (3):

Among them, NNI _v (u) is the relative influence value of the neighboring node v of node u on node u; B(c _v ,v) is the attribution degree of node v to the community number c _v ; c is the value in L _Ng A community number; in the calculation, all the attribution degrees in the first label set participate in the denominator calculation, and the attribution degrees with the community number c in the first label set participate in the numerator calculation; B′(c,u) is the pair of node u The degree of attribution of the community whose community number is c in the first label set.

After calculating the attribution degree B′(c,u) of the node to each community number in the first label set, combine the attribution degree B′(c,u) and its corresponding community number to obtain multiple community tags , Obtain the second label set of the node according to multiple labels.

In one embodiment, after calculating a node according to formula (3), a new label set L′ _{Ng is obtained} (using |L′| to represent the _{number of new community labels in L Ng} ′, and |L′| is equal to The number of community tags in the first tag set is equal to the number of neighboring nodes of node u):

The attribution degree of some community tags in L′ _Ng may be less than 1/|L′|. Such attribution is too low. You can delete the community tag where it belongs from |L′| to get the community tag set L″ _Ng . At this time _{, the sum of all attribution degrees in L″Ng} is not equal to 1, and normalization is performed:

The normalized community label set L″ _{Ng is} used as the second label set of the node.

Step 2054: From the second tag set, filter at least one community tag whose attribution degree meets the preset attribution degree condition.

Wherein, the preset attribution degree condition may be a preset condition for screening attribution degree.

Specifically, the server obtains the preset attribution degree condition, searches for the attribution degree that meets the preset attribution degree condition from the second set of tags according to the preset attribution degree condition, and filters the community tag where the found attribution degree is located. The server can filter at least one community tag from the second tag set according to the preset attribution degree condition.

In one embodiment, the server searches for the maximum attribution degree, obtains the preset interval adjustment value, uses the maximum attribution degree as the right end of the interval, and uses the difference between the maximum attribution degree and the preset interval adjustment value as the left end of the interval to obtain a closed interval . The server screens the community tags whose attribution is within the closed interval.

For example, there are 3 community tags in the second tag set, and their attribution degrees are 0.6, 0.3, and 0.1 in order. Then 0.6 is the maximum attribution degree, the preset interval adjustment value is 0.3, then the left end of the interval is 0.6-0.3=0.3, and the closed interval [0.3, 0.6] is obtained. Then, two community tags are screened out, which are the community tags where the attribution degree is 0.6 and 0.3.

Step 2055: Use the at least one community label selected as the community label set of the node after the iterative update.

Specifically, the server obtains the community label set according to at least one community label that is screened out, and the newly obtained community label set is the community label set after the node is iteratively updated in this round.

In this embodiment, the nodes are sorted in ascending order according to the node influence value, and the iterative update is performed according to the sort, which ensures that the node with the largest node influence value can have the greatest influence. When updating each node, obtain the relative influence value of each adjacent node of the node on the node and the first label set of the node. The first label set is the community label with the largest degree of belonging in the community label set of each adjacent node ; Based on the relative influence value, calculate the attribution degree of the node to the community number in each community label in the first label set to obtain the second label set. The calculation of the second label set fully considers the influence of neighboring nodes; from the second label set At least one community label whose attribution degree meets the preset attribution degree condition is selected, so as to obtain the community label set after the node iteration, which ensures the accuracy of the updated community label set.

Further, as shown in FIG. 7, the foregoing step 206 may include:

Step 2061: After completing each round of iterative update for each node, determine the community label with the largest attribution in the community label set of each node as the main label of the community label set of each node.

Specifically, after the server completes a round of iterative update of each node in the homogeneous network, it compares the attribution degree of the community tag in the community tag set of each node, finds the maximum attribution degree, and takes the community tag where the maximum attribution degree is located as The main label in the community label set of each node.

Step 2062: Compare the current primary label of each node with whether the primary label of each node has changed after the previous iteration.

Specifically, the server obtains the primary label of each node after the previous iteration update, and compares whether the current primary label has changed with the primary label after the previous iteration update. Wherein, the change includes at least one of a change in the community number of the main tag and a change in the degree of attribution.

Step 2063: If there is no change, divide the homogenous network into at least one community node network according to the community number of the current community tag of each node.

Specifically, if the primary label of each node after the current iteration update has not changed compared with the primary label after the previous iteration update, the operation on the homogeneous network has converged, and the homogeneous network has stabilized. The server reads the community number in the current community label set of each node, and divides the nodes with the same community number into a community node network. The server divides at least one community node network from the homogeneous network. If the main label of each node after the current iteration update has changed compared with the main label after the previous iteration update, the iterative update is continued.

A node can be divided into multiple community node networks. When there is more than one community label in the community label of a node, the node is divided into more than one community node network and becomes an overlapping node.

In practical applications, transaction users who have close connections in financial transactions can be divided into the same community node network through the embodiments of this application. For example, a homogeneous network containing fraudulent network groups can be stable after a certain number of iterations and updates, and nodes represented by fraudulent groups are classified into a community node network.

In this embodiment, after each round of iterative update is completed, the main tags are filtered from the community tag set of each node; if the current main tag of each node has not changed compared with the main tag after the previous iteration, it is homogeneous The network tends to be stable. According to the community number of the current community tag of each node, the community node network can be accurately divided from the homogeneous network.

Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through computer-readable instructions, which can be stored in a computer-readable storage medium. When the computer-readable instructions are executed, they may include the processes of the above-mentioned method embodiments.

Although the steps in the flowchart of the drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless explicitly stated in this article, the execution of these steps is not strictly limited in order, and they can be executed in other orders.

With further reference to FIG. 8, as an implementation of the method shown in FIG. 2, this application provides an embodiment of a device for community division of a homogeneous network, and the device embodiment corresponds to the method embodiment shown in FIG. 2 , The device can be specifically applied to various electronic equipment.

As shown in FIG. 4, the device 400 for community division of a homogeneous network in this embodiment includes: an information receiving module 401, a network construction module 402, a label initial module 403, a node calculation module 404, a label update module 405, and network division Module 406. in:

The information receiving module 401 is used to receive financial transaction information.

The network construction module 402 is configured to use transaction users in the financial transaction information as nodes, and construct a homogeneous network based on transaction relationships and transaction data between the transaction users.

The label initialization module 403 is used to initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network.

The node calculation module 404 is configured to calculate each node using transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on adjacent nodes.

The label update module 405 is configured to iteratively update the community label set of each node according to the node influence value and the relative influence value; there is at least one community label in the community label set after the iterative update.

The network division module 406 is used to divide the homogeneous network into at least one community node network according to the community label after the iterative update of each node when the community label in the community label set of each node no longer changes .

In this embodiment, when calculating the node influence value and relative influence value of each node, the transaction data is used as the weight, taking into account the importance of the node, and improving the accuracy of the calculation; a node can be divided into at least one community node network , To ensure the accuracy of homogenous network community division.

In some optional implementations of this embodiment, the above-mentioned network construction module 402 is further used to identify transaction users in the financial transaction information and transaction relationships and transaction data between the transaction users; and use the identified transaction users as the to-be-built Nodes in a homogeneous network; nodes that have a transaction relationship are connected through a node connection line, and the transaction data corresponding to the transaction relationship is marked on the node connection line to obtain a homogenous network.

In this embodiment, the identified transaction users are used as nodes, nodes with transaction relationships are connected through node connection lines, and the corresponding transaction data is marked on the node connection lines, ensuring the accuracy of constructing a homogeneous network.

In some optional implementations of this embodiment, the above-mentioned label initializing module 403 is further configured to configure community numbers for each node in the homogeneous network; initialize the attribution degree of each node to the configured community number; The community number of the node and the degree of belonging corresponding to the community number generate the community label of each node; and the community label set of each node is constructed according to the community label of each node.

In this embodiment, the community number and corresponding attribution of each node are initialized, and a community label and a community label set are generated, which ensures the normal progress of subsequent community division.

In some optional implementations of this embodiment, the above-mentioned node calculation module 404 is further configured to use transaction data as a weight, and calculate the transaction weight value of each node according to the transaction relationship; for each node in the homogeneous network, the slave node's Among the transaction weight values corresponding to adjacent nodes, filter the maximum transaction weight value and the minimum transaction weight value; according to the transaction weight value, maximum transaction weight value and minimum transaction weight value of the node, calculate the node influence value of the node in the homogeneous network .

In this embodiment, when performing node analysis and calculating the node influence value, the transaction data is substituted into the calculation as the weight, which takes into account the difference in financial transactions between the users corresponding to the node, and improves the accuracy of the calculated node influence value. .

In some optional implementations of this embodiment, the above-mentioned node calculation module 404 is further configured to calculate the similarity between each pair of adjacent nodes according to the transaction relationship and transaction data; for each node in the homogeneous network, the slave node The maximum similarity of adjacent nodes is determined from the similarity corresponding to adjacent nodes; based on the node influence value of the node, the similarity between the node and the adjacent node, and the maximum similarity of the adjacent node, the calculation of the node-to-adjacent node The relative influence value of.

In this embodiment, the relative influence value is calculated based on the node influence value, and the calculation of the node influence value uses transaction data as a weight, which has high accuracy, thereby ensuring the accuracy of the relative influence value.

In some optional implementations of this embodiment, the label update module 405 is further configured to sort the nodes in ascending order according to the node influence value to obtain the update sequence sequence; for each node in the update sequence sequence, obtain the node in sequence The relative influence value of each neighboring node on the node and the first label set of the node; the first label set is composed of the community label with the largest attribution in the community label set of each neighboring node; based on the relative influence value, the node pair is calculated The attribution degree of the community number in each community tag in the first tag set is obtained, and the second tag set is obtained; from the second tag set, at least one community tag whose attribution degree meets the preset attribution degree condition is selected; The group label is used as the community label set of the node after iterative update.

In this embodiment, the update sequence of the community label takes into account the node influence value, and the update method of the community label takes the relative influence value into consideration, thereby ensuring the accuracy of updating the community label set.

In some optional implementations of this embodiment, the above-mentioned network division module 406 is further configured to collect the community labels of each node with the community label with the largest degree of attribution after completing each round of iterative update of each node, and determine Is the main label of the community label set of each node; compare the current main label of each node with the main label of each node after the previous iteration; if there is no change, according to the community number of the current community label of each node, Divide the homogeneous network into at least one community node network.

In this embodiment, the homogeneous network tends to be stable when the primary label of each node no longer changes, and the community node network can be accurately divided according to the community number of the current community label.

In order to solve the above technical problems, the embodiments of the present application also provide computer equipment. Please refer to FIG. 9 for details. FIG. 9 is a block diagram of the basic structure of the computer device in this embodiment.

The computer device 5 includes a memory 51, a processor 52, and a network interface 53 that communicate with each other through a system bus. The figure only shows the computer device 5 with components 51-53, and it is not required to implement all the illustrated components, and more or fewer components may be implemented instead. .

The memory 51 includes at least one type of computer-readable storage medium. The computer-readable storage medium may be nonvolatile or volatile. The computer-readable storage medium includes flash memory, hard disk, and multimedia card. , Card-type memory, random access memory, static random access memory, read-only memory, electrically erasable programmable read-only memory, programmable read-only memory, magnetic memory, magnetic disk, optical disk, etc. The memory 51 may be an internal storage unit of the computer device 5, such as a hard disk or a memory, or an external storage device of the computer device 5, such as a plug-in hard disk, a smart memory card, a secure digital card, or a flash memory card. Wait. Of course, the memory 51 may also include both the internal storage unit of the computer device 5 and its external storage device. In this embodiment, the memory 51 is generally used to store an operating system and various application software installed in the computer device 5, such as computer-readable instructions for a community division method of a homogeneous network. In addition, the memory 51 can also be used to temporarily store various types of data that have been output or will be output.

The processor 52 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 52 is generally used to control the overall operation of the computer device 5. In this embodiment, the processor 52 is configured to run computer-readable instructions or process data stored in the memory 51, for example, computer-readable instructions for running the community division method of the homogeneous network.

The network interface 53 may include a wireless network interface or a wired network interface, and the network interface 53 is generally used to establish a communication connection between the computer device 5 and other electronic devices.

The computer device provided in this embodiment can execute the steps of the above-mentioned method for community division of a homogeneous network. Here, the steps of the method for dividing a community of a homogeneous network may be the steps of the method for dividing a community of a homogeneous network in each of the foregoing embodiments.

In this embodiment, when calculating the node influence value and relative influence value of each node, the transaction data is used as the weight, taking into account the importance of the node, and improving the accuracy of the calculation; a node can be divided into at least one community node network , To ensure the accuracy of homogenous network community division.

This application also provides another implementation manner, that is, a computer-readable storage medium that stores computer-readable instructions for community division of a homogenous network. The computer-readable instructions for the community division of the network may be executed by at least one processor, so that the at least one processor executes the steps of the community division method of the homogeneous network as described above.

In this embodiment, when calculating the node influence value and relative influence value of each node, the transaction data is used as the weight, taking into account the importance of the node, and improving the accuracy of the calculation; a node can be divided into at least one community node network , To ensure the accuracy of homogenous network community division.

Through the description of the above implementation manners, those skilled in the art can clearly understand that the above-mentioned embodiment method can be implemented by means of software plus the necessary general hardware platform, of course, it can also be implemented by hardware, but in many cases the former is better.的实施方式。 Based on this understanding, the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM/RAM, magnetic disk, The optical disc) includes several instructions to make a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of the present application.

Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The drawings show preferred embodiments of the present application, but do not limit the patent scope of the present application. This application can be implemented in many different forms. Any equivalent structure made using the content of the description and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims (20)

1. A method of community division in a homogeneous network includes the following steps:

   Receive financial transaction information;

   Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

   Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

   Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

   According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

   When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.
2. The method for community division of a homogenous network according to claim 1, wherein the transaction users in the financial transaction information are used as nodes, and the homogenous network is constructed based on the transaction relationship and transaction data between the transaction users The steps include:

   Identifying the transaction user in the financial transaction information and the transaction relationship and transaction data between the transaction users;

   Use the identified transaction users as nodes in the homogeneous network to be constructed;

   The nodes with transaction relationships are connected through node connecting lines, and transaction data corresponding to the transaction relationships are marked on the node connecting lines to obtain a homogeneous network.
3. The method for dividing a community in a homogeneous network according to claim 1, wherein the step of initializing the community label of each node in the homogenous network to obtain the community label set of each node in the homogeneous network include:

   Respectively configuring a community number to each node in the homogeneous network;

   Initialize the attribution degree of each node to the configured community number;

   Generating the community label of each node based on the community number of each node and the attribution degree corresponding to the community number;

   The community label set of each node is respectively constructed according to the community label of each node.
4. The method for community division of a homogeneous network according to claim 1, wherein the step of calculating each node using the transaction data as a weight to obtain the node influence value of each node comprises:

   Using the transaction data as a weight, and calculating the transaction weight value of each node according to the transaction relationship;

   For each node in the homogeneous network, filter the maximum transaction weight value and the minimum transaction weight value from the transaction weight values corresponding to the neighboring nodes of the node;

   According to the transaction weight value of the node, the maximum transaction weight value and the minimum transaction weight value, the node influence value of the node in the homogeneous network is calculated.
5. The method for dividing a community in a homogeneous network according to claim 4, wherein the step of calculating the relative influence value of each node on neighboring nodes comprises:

   According to the transaction relationship and transaction data, calculate the similarity between each pair of adjacent nodes;

   For each node in the homogeneous network, determine the maximum similarity of adjacent nodes from the similarity corresponding to the adjacent nodes of the node;

   Based on the node influence value of the node, the similarity between the node and the adjacent node, and the maximum similarity of the adjacent node, the relative influence value of the node on the adjacent node is calculated.
6. The method for community division of a homogeneous network according to claim 3, wherein the step of iteratively updating the community label set of each node according to the node influence value and the relative influence value comprises:

   According to the node influence value, sort the nodes in ascending order to obtain an update sequence sequence;

   For each node in the update sequence sequence, the relative influence value of each adjacent node of the node on the node and the first label set of the node are sequentially obtained; the first label set is determined by the adjacent nodes The community label set is composed of the community label with the greatest degree of attribution;

   Based on the relative influence value, calculating the attribution degree of the node to the community number in each community label in the first label set to obtain a second label set;

   From the second tag set, filter at least one community tag whose attribution degree meets the preset attribution degree condition;

   At least one community label selected is used as the community label set of the node after iterative update.
7. The method for community division of a homogeneous network according to claim 3, wherein, when the community label in the community label set of each node no longer changes in the iterative update, iteratively update according to each node The step of dividing the homogeneous network into at least one community node network by the following community label includes:

   After completing each round of iterative update for each node, determine the community label with the largest attribution in the community label set of each node as the main label of the community label set of each node;

   Compare the current primary label of each node with the primary label of each node after the previous iteration;

   If there is no change, divide the homogenous network into at least one community node network according to the community number of the current community label of each node.
8. A device for dividing a homogeneous network into a community, including:

   Information receiving module for receiving financial transaction information;

   The network construction module is configured to use the transaction users in the financial transaction information as nodes, and construct a homogeneous network based on the transaction relationship and transaction data between the transaction users;

   The label initialization module is used to initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

   A node calculation module, configured to calculate each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

   The label update module is configured to iteratively update the community label set of each node according to the node influence value and the relative influence value; there is at least one community label in the community label set after the iterative update;

   The network division module is used to divide the homogenous network into at least the same network according to the community label after the iterative update of each node when the community label in the community label set of each node no longer changes. A network of community nodes.
9. A computer device includes a memory and a processor. The memory stores computer readable instructions. When the processor executes the computer readable instructions, the following steps are implemented:

   Receive financial transaction information;

   Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

   Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

   Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

   According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

   When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.
10. The computer device according to claim 9, wherein the step of initializing the community label of each node in the homogenous network to obtain the community label set of each node in the homogenous network comprises:

    Respectively configuring a community number to each node in the homogeneous network;

    Initialize the attribution degree of each node to the configured community number;

    Generating the community label of each node based on the community number of each node and the attribution degree corresponding to the community number;

    The community label set of each node is respectively constructed according to the community label of each node.
11. The computer device according to claim 9, wherein the step of calculating each node using the transaction data as a weight to obtain the node influence value of each node comprises:

    Using the transaction data as a weight, and calculating the transaction weight value of each node according to the transaction relationship;

    For each node in the homogeneous network, filter the maximum transaction weight value and the minimum transaction weight value from the transaction weight values corresponding to the neighboring nodes of the node;

    According to the transaction weight value of the node, the maximum transaction weight value and the minimum transaction weight value, the node influence value of the node in the homogeneous network is calculated.
12. The computer device according to claim 11, wherein the step of calculating the relative influence value of each node on neighboring nodes comprises:

    According to the transaction relationship and transaction data, calculate the similarity between each pair of adjacent nodes;

    For each node in the homogeneous network, determine the maximum similarity of adjacent nodes from the similarity corresponding to the adjacent nodes of the node;

    Based on the node influence value of the node, the similarity between the node and the adjacent node, and the maximum similarity of the adjacent node, the relative influence value of the node on the adjacent node is calculated.
13. 10. The computer device according to claim 10, wherein the step of iteratively updating the community tag set of each node according to the node influence value and the relative influence value comprises:

    According to the node influence value, sort the nodes in ascending order to obtain an update sequence sequence;

    For each node in the update sequence sequence, the relative influence value of each adjacent node of the node on the node and the first label set of the node are sequentially obtained; the first label set is determined by the adjacent nodes The community label set is composed of the community label with the greatest degree of attribution;

    Based on the relative influence value, calculate the attribution degree of the node to the community number in each community label in the first label set to obtain a second label set;

    From the second tag set, filter at least one community tag whose attribution degree meets the preset attribution degree condition;

    At least one community label selected is used as the community label set of the node after iterative update.
14. The computer device according to claim 10, wherein when the community label in the community label set of each node no longer changes in the iterative update, the community label after the iterative update of each node is changed The step of dividing the homogeneous network into at least one community node network includes:

    After completing each round of iterative update for each node, determine the community label with the largest attribution in the community label set of each node as the main label of the community label set of each node;

    Compare the current primary label of each node with the primary label of each node after the previous iteration;

    If there is no change, divide the homogenous network into at least one community node network according to the community number of the current community label of each node.
15. A computer-readable storage medium having computer-readable instructions stored thereon, and when the computer-readable instructions are executed by a processor, the following steps are implemented:

    Receive financial transaction information;

    Using transaction users in the financial transaction information as nodes, and constructing a homogeneous network based on the transaction relationship and transaction data between the transaction users;

    Initialize the community label of each node in the homogeneous network to obtain the community label set of each node in the homogeneous network;

    Calculating each node using the transaction data as a weight to obtain the node influence value of each node and the relative influence value of each node on neighboring nodes;

    According to the node influence value and the relative influence value, iteratively update the community label set of each node; there is at least one community label in the community label set after the iterative update;

    When the community label in the community label set of each node no longer changes during the iterative update, the homogeneous network is divided into at least one community node network according to the community label after the iterative update of each node.
16. 15. The computer-readable storage medium according to claim 15, wherein the step of initializing the community label of each node in the homogenous network to obtain the community label set of each node in the homogenous network comprises:

    Respectively configuring a community number to each node in the homogeneous network;

    Initialize the attribution degree of each node to the configured community number;

    Generating the community label of each node based on the community number of each node and the attribution degree corresponding to the community number;

    The community label set of each node is respectively constructed according to the community label of each node.
17. The computer-readable storage medium according to claim 15, wherein the step of calculating each node using the transaction data as a weight to obtain the node influence value of each node comprises:

    Using the transaction data as a weight, and calculating the transaction weight value of each node according to the transaction relationship;

    For each node in the homogeneous network, filter the maximum transaction weight value and the minimum transaction weight value from the transaction weight values corresponding to the neighboring nodes of the node;

    According to the transaction weight value of the node, the maximum transaction weight value and the minimum transaction weight value, the node influence value of the node in the homogeneous network is calculated.
18. 18. The computer-readable storage medium according to claim 17, wherein the step of calculating the relative influence value of each node on neighboring nodes comprises:

    According to the transaction relationship and transaction data, calculate the similarity between each pair of adjacent nodes;

    For each node in the homogeneous network, determine the maximum similarity of adjacent nodes from the similarity corresponding to the adjacent nodes of the node;

    Based on the node influence value of the node, the similarity between the node and the adjacent node, and the maximum similarity of the adjacent node, the relative influence value of the node on the adjacent node is calculated.
19. 16. The computer-readable storage medium according to claim 16, wherein the step of iteratively updating the community tag set of each node according to the node influence value and the relative influence value comprises:

    According to the node influence value, sort the nodes in ascending order to obtain an update sequence sequence;

    For each node in the update sequence sequence, the relative influence value of each adjacent node of the node on the node and the first label set of the node are sequentially obtained; the first label set is determined by the adjacent nodes The community label set is composed of the community label with the greatest degree of attribution;

    Based on the relative influence value, calculating the attribution degree of the node to the community number in each community label in the first label set to obtain a second label set;

    From the second tag set, filter at least one community tag whose attribution degree meets the preset attribution degree condition;

    At least one community label selected is used as the community label set of the node after iterative update.
20. The computer-readable storage medium according to claim 16, wherein, when the community label in the community label set of each node no longer changes during the iterative update, according to the community label after the iterative update of each node The step of dividing the homogeneous network into at least one community node network by the group label includes:

    After completing each round of iterative update for each node, determine the community label with the largest attribution in the community label set of each node as the main label of the community label set of each node;

    Compare the current primary label of each node with the primary label of each node after the previous iteration;

    If there is no change, divide the homogenous network into at least one community node network according to the community number of the current community label of each node.

Images