Risk assessment method and device for financial account, electronic equipment and storage medium

Abstract

The invention discloses a risk assessment method and device for a financial account, electronic equipment and a storage medium, and the method comprises the steps: collecting account data of N customer accounts in a preset time period, obtaining an account data set, determining a community set to which each node belongs based on the account data set, and carrying out the calculation of the community set, the method comprises the steps of obtaining an account data set, dividing nodes into community sets to which the nodes belong, obtaining M community sets, determining preset nodes in each community set, determining relation values between other nodes in the community sets and the preset nodes on the basis of the account data set, and determining a risk coefficient of each node in the community sets on the basis of the relation values corresponding to the other nodes, and when the target risk coefficient of the target node is greater than a preset risk threshold, determining that the target customer account associated with the target risk coefficient has a risk. The technical problem that the risk of the financial institution is easily caused because effective risk assessment cannot be carried out on the customer account of the financial institution in the related technology is solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a risk assessment method for a financial account and a device, electronic equipment and storage medium thereof. Background Art

[0002] Risk control management is an important part of the business operations of financial institutions. Through credit assessment and risk control of customers, the safety of financial institutions' assets can be guaranteed and the stability of financial markets can be maintained. However, financial institutions have a large customer base and a large amount of customer information. How to conduct risk assessment and control of customers is a problem that needs to be solved.

[0003] Related technologies often use data mining technology to analyze customers' transaction behaviors, credit behaviors, etc. based on their historical data to assess their own risks and then conduct risk prevention and control.

[0004] However, the risk assessment methods in related technologies have the following problems: (1) The data source requirements are high. The disadvantage of existing machine learning models is that they require a large amount of historical data for training, and new risk types may require retraining of the model, so data and algorithms need to be continuously updated; (2) The analysis results are limited. The related technology is only based on the financial institution’s own transaction data and credit assessment data for analysis, and cannot conduct in-depth analysis of the customer’s social network relationships, and cannot fully discover the customer’s potential risk factors.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present invention provide a risk assessment method and device for a financial account, an electronic device, and a storage medium, so as to at least solve the technical problem in the related art that it is impossible to conduct effective risk assessment on customer accounts of financial institutions, which easily leads to risks for financial institutions.

[0007] According to one aspect of an embodiment of the present invention, a risk assessment method for a financial account is provided, comprising: collecting account data of N customer accounts within a preset time period to obtain an account data set, wherein N is a positive integer and each customer account is characterized as a node; based on the account data set, determining a community set to which each node belongs, and dividing the nodes into the community sets to which they belong, to obtain M community sets, wherein M is a positive integer; determining a preset node in each of the community sets; based on the account data set, determining a relationship value between other nodes in the community set and the preset node, and determining a risk coefficient of each node in the community set based on the relationship values ​​corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset nodes; and when a target risk coefficient of a target node is greater than a preset risk threshold, determining that a target customer account associated with the target risk coefficient is at risk.

[0008] Optionally, the step of determining the community set to which each of the nodes belongs based on the account data set includes: characterizing each of the nodes as an initial community set; traversing all of the nodes, and determining a first modularity increment value of a first adjacent community set after the traversed first node is placed into the first adjacent community set, wherein the first adjacent community set is the community set in which the node distance between the first node and all first adjacent nodes in the first adjacent community set is less than a preset distance threshold, and the first adjacent node is a node in the first adjacent community set; when the first modularity increment value is greater than a preset increment threshold, determining to move the first node into the first adjacent community set, and generating multiple preset community sets when all of the nodes are traversed; based on the account data set, updating each of the preset community sets until the modularity of all of the preset community sets no longer increases, thereby obtaining multiple target community sets; and determining the community set to which each of the nodes belongs based on the multiple target community sets.

[0009] Optionally, the step of determining a first modularity increment value of the first neighboring community set after the traversed first node is placed into the first neighboring community set includes: determining all the first neighboring nodes included in the first neighboring community set before the traversed first node is placed into the first neighboring community set; calculating the first modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes; after the traversed first node is placed into the first neighboring community set, calculating the second modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes and the account data associated with the first node; and calculating the difference between the second modularity and the first modularity to obtain the first modularity increment value.

[0010] Optionally, based on the account data set, updating each of the preset community sets until the modularity of all the preset community sets no longer increases, and obtaining a plurality of target community sets comprises: traversing the nodes in all the preset community sets, and determining a second modularity increment value of the second adjacent community set after the traversed second node is placed into the second adjacent community set, wherein the second adjacent community set is the preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than the preset distance threshold, and the second adjacent node is a node in the second adjacent community set; when the second modularity increment value is greater than the preset increment threshold, determining to move the second node into the second adjacent community set; and continuing to determine a third modularity increment value of the third adjacent community set after the traversed third node is placed into the third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, and obtaining a plurality of target community sets.

[0011] Optionally, the step of determining the preset nodes in each of the community sets includes: determining a first center value of each of the nodes in the community set, and selecting the nodes indicated by the first center value greater than a first preset center threshold to obtain a first node set; determining a second center value of each of the nodes in the community set, and selecting the nodes indicated by the second center value greater than a second preset center threshold to obtain a second node set; determining a third center value of each of the nodes in the community set, and selecting the nodes indicated by the third center value greater than a third preset center threshold to obtain a third node set; performing an intersection operation on the first node set, the second node set and the third node set to obtain a fourth node set; when the fourth node set includes multiple nodes, selecting the node indicated by the maximum average center value as the preset node in the community set, wherein the average center value is the average value between the first center value, the second center value and the third center value.

[0012] Optionally, the risk assessment method also includes: determining the number of nodes in the community set and the node degrees of the nodes in the community set, and determining the first center value of the node based on the number of nodes and the node degrees; determining the shortest path distance between the node and the remaining nodes in the community set, and determining the second center value of the node based on the number of nodes and all the shortest path distances, wherein the remaining nodes are nodes in the community set other than the node; determining the number of paths in which the shortest path between any two nodes in the community set passes through the node, and determining the third center value of the node based on the number of nodes and the number of paths.

[0013] Optionally, the step of determining the relationship value between other nodes in the community set and the preset node based on the account data set includes: obtaining transaction data between the other nodes and the preset node from the account data set; and determining the relationship value based on the transaction data.

[0014] Optionally, the step of determining the risk coefficient of each of the nodes in the community set based on the relationship values ​​corresponding to the other nodes includes: determining a first coefficient for each of the nodes in the community set; screening the relationship values ​​greater than a preset relationship threshold to obtain a target relationship value set, and determining a second coefficient corresponding to each target relationship value in the target relationship value set; when the node corresponds to the second coefficient, accumulating the first coefficient and the second coefficient to obtain the risk coefficient of the node; or, when the node does not have the second coefficient, characterizing the first coefficient as the risk coefficient of the node.

[0015] Optionally, after collecting account data of N customer accounts within a preset time period and obtaining an account data set, the method further includes: determining a preset number of clusters, and selecting K cluster centers from all the nodes, wherein K is a positive integer, K is equal to the preset number of clusters, and the cluster center corresponds to a cluster set; determining the cluster set to which each of the nodes belongs, wherein the step of determining the cluster set to which each of the nodes belongs includes: determining the distance between the node and the cluster center, and adding the node to the cluster set where the cluster center indicated by the minimum distance is located; when determining the cluster set to which all of the nodes belong, updating the cluster centers in the cluster set, and executing the step of determining the cluster set to which each of the nodes belongs until the cluster center in each of the cluster sets no longer changes or the number of executions is reached.

[0016] According to another aspect of an embodiment of the present invention, a risk assessment device for a financial account is provided, comprising: a collection unit, configured to collect account data of N customer accounts within a preset time period to obtain an account data set, wherein N is a positive integer and each customer account is characterized as a node; a first determination unit, configured to determine, based on the account data set, a community set to which each of the nodes belongs, and to divide the nodes into the community sets to which they belong, to obtain M community sets, wherein M is a positive integer; a second determination unit, configured to determine a preset node in each of the community sets; a third determination unit, configured to determine, based on the account data set, a relationship value between other nodes in the community set and the preset node, and to determine a risk coefficient of each of the nodes in the community set based on the relationship values ​​corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset nodes; and a fourth determination unit, configured to determine, when a target risk coefficient of a target node is greater than a preset risk threshold, that a target customer account associated with the target risk coefficient is at risk.

[0017] Optionally, the first determination unit includes: a first characterization module, configured to characterize each of the nodes as an initial community set; a first traversal module, configured to traverse all of the nodes and determine a first modularity increment value of the first adjacent community set after the traversed first node is placed into the first adjacent community set, wherein the first adjacent community set is the community set in which the node distance between the first node and all first adjacent nodes in the first adjacent community set is less than a preset distance threshold, and the first adjacent node is a node in the first adjacent community set; a first determination module, configured to determine to move the first node into the first adjacent community set when the first modularity increment value is greater than a preset increment threshold, and to generate multiple preset community sets when all of the nodes are traversed; a first update module, configured to update each of the preset community sets based on the account data set until the modularity of all the preset community sets no longer increases, thereby obtaining multiple target community sets; and a second determination module, configured to determine the community set to which each of the nodes belongs based on the multiple target community sets.

[0018] Optionally, the first traversal module includes: a first determination submodule, used to determine all the first neighboring nodes included in the first neighboring community set before placing the traversed first node into the first neighboring community set; a first calculation submodule, used to calculate the first modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes; a second calculation submodule, used to calculate the second modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes and the account data associated with the first node after placing the traversed first node into the first neighboring community set; and a third calculation submodule, used to calculate the difference between the second modularity and the first modularity to obtain the first modularity increment value.

[0019] Optionally, the first updating module includes: a first traversal submodule, used to traverse the nodes in all the preset community sets, and determine a second modularity increment value of the second adjacent community set after the traversed second node is placed into the second adjacent community set, wherein the second adjacent community set is the preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than the preset distance threshold, and the second adjacent node is a node in the second adjacent community set; a second determination submodule, used to determine to move the second node into the second adjacent community set when the second modularity increment value is greater than the preset increment threshold; and a third determination submodule, used to continue to determine a third modularity increment value of the third adjacent community set after the traversed third node is placed into the third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, thereby obtaining multiple target community sets.

[0020] Optionally, the second determination unit includes: a third determination module, used to determine the first center value of each of the nodes in the community set, and select the nodes indicated by the first center value greater than a first preset center threshold to obtain a first node set; a fourth determination module, used to determine the second center value of each of the nodes in the community set, and select the nodes indicated by the second center value greater than a second preset center threshold to obtain a second node set; a fifth determination module, used to determine the third center value of each of the nodes in the community set, and select the nodes indicated by the third center value greater than a third preset center threshold to obtain a third node set; a first operation module, used to perform an intersection operation on the first node set, the second node set and the third node set to obtain a fourth node set; a first selection module, used to select the node indicated by the maximum average center value as the preset node in the community set when the fourth node set includes multiple nodes, wherein the average center value is the average value between the first center value, the second center value and the third center value.

[0021] Optionally, the risk assessment device also includes: a sixth determination module, used to determine the number of nodes in the community set and the node degree of the nodes in the community set, and determine the first center value of the node based on the number of nodes and the node degree; a seventh determination module, used to determine the shortest path distance between the node and the remaining nodes in the community set, and determine the second center value of the node based on the number of nodes and all the shortest path distances, wherein the remaining nodes are nodes in the community set other than the node; an eighth determination module, used to determine the number of paths in which the shortest path between any two nodes in the community set passes through the node, and determine the third center value of the node based on the number of nodes and the number of paths.

[0022] Optionally, the third determination unit includes: a first acquisition module, used to acquire transaction data between the other nodes and the preset node from the account data set; and a ninth determination module, used to determine the relationship value based on the transaction data.

[0023] Optionally, the third determination unit also includes: a tenth determination module, used to determine the first coefficient of each of the nodes in the community set; a first screening module, used to screen the relationship values ​​greater than a preset relationship threshold, obtain a target relationship value set, and determine the second coefficient corresponding to each target relationship value in the target relationship value set; a first accumulation module, used to accumulate the first coefficient and the second coefficient to obtain the risk coefficient of the node when the node corresponds to the second coefficient; and a second characterization module, used to characterize the first coefficient as the risk coefficient of the node when the node does not have the second coefficient.

[0024] Optionally, the risk assessment device also includes: an eleventh determination module, which is used to determine the preset number of clusters after collecting account data of N customer accounts within a preset time period and obtaining an account data set, and select K cluster centers from all the nodes, wherein K is a positive integer, K is equal to the preset number of clusters, and the cluster center corresponds to a cluster set; a twelfth determination module, which is used to determine the cluster set to which each of the nodes belongs, wherein the step of determining the cluster set to which each of the nodes belongs includes: determining the distance between the node and the cluster center, and adding the node to the cluster set where the cluster center indicated by the minimum distance is located; a second updating module, which is used to update the cluster centers in the cluster set when determining the cluster set to which all the nodes belong, and execute the step of determining the cluster set to which each of the nodes belongs until the cluster center in each of the cluster sets no longer changes or the number of executions is reached.

[0025] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned financial account risk assessment methods.

[0026] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned risk assessment methods for a financial account.

[0027] In the present disclosure, account data of N customer accounts within a preset time period are collected to obtain an account data set. Based on the account data set, the community set to which each node belongs is determined, and the nodes are divided into the community sets to which they belong to obtain M community sets. The preset nodes in each community set are determined. Based on the account data set, the relationship values ​​between other nodes in the community set and the preset nodes are determined, and the risk coefficient of each node in the community set is determined based on the relationship values ​​corresponding to the other nodes. When the target risk coefficient of the target node is greater than the preset risk threshold, it is determined that the target customer account associated with the target risk coefficient is at risk. In the present disclosure, nodes representing customer accounts can be allocated to corresponding community sets based on the collected account data set, and then the preset nodes in each community set are determined. Then, the relationship values ​​between other nodes in each community set and the preset nodes are determined based on the account data set, and the risk coefficient of each node in the community set is determined based on the relationship values ​​corresponding to other nodes. If there is a node whose risk coefficient is greater than a preset risk threshold, it is determined that the corresponding customer account is at risk, and the risk is handled in a timely manner. By deeply mining the relationship between customer accounts, the accuracy of risk assessment of customer accounts can be effectively improved, providing a more comprehensive reference basis for risk control and management of financial institutions, thereby solving the technical problem that the relevant technology cannot effectively assess the risk of customer accounts of financial institutions, which easily leads to risks for financial institutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0029] Figure 1 is a flow chart of an optional risk assessment method for a financial account according to an embodiment of the present invention;

[0030] Figure 2is a schematic diagram of an optional risk assessment device for a financial account according to an embodiment of the present invention;

[0031] Figure 3 The present invention is a hardware structure block diagram of an electronic device (or mobile device) for a risk assessment method for a financial account according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0035] In the present invention, social network analysis technology is a technology that can mine associations and group behaviors in data. By analyzing the customer relationship network, the associations and potential risk factors between customers can be discovered, providing a reference for risk control and management of financial institutions.

[0036] The present invention can construct a risk customer association map based on social network analysis technology. The purpose of the present invention is to provide a new customer risk assessment method for financial institutions. The customer risk assessment method in the related art mainly relies on the collection and analysis of a single customer's information, and cannot fully understand the customer's risk situation. The present invention can more fully understand the customer's risk situation by analyzing the relationship and interaction between customers, and can help financial institutions identify individual customers who may pose potential risks to the entire customer group.

[0037] Therefore, the present invention can integrate internal and external data resources of financial institutions, use social network analysis technology to construct a risk customer association map, and then conduct in-depth mining of customer relationships, discover the relationships between customers and potential risk factors, and provide a more comprehensive reference basis for risk control and management of financial institutions.

[0038] The present invention is described in detail below in conjunction with various embodiments.

[0039] Embodiment 1

[0040] According to an embodiment of the present invention, an embodiment of a risk assessment method for a financial account is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0041] Figure 1 is a flow chart of an optional financial account risk assessment method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0042] Step S101, collecting account data of N customer accounts within a preset time period to obtain an account data set, wherein N is a positive integer and each customer account is represented as a node.

[0043] Step S102: based on the account data set, determine the community set to which each node belongs, and divide the nodes into the community sets to which they belong, to obtain M community sets, where M is a positive integer.

[0044] Step S103: determining the preset nodes in each community set.

[0045] Step S104, based on the account data set, determining the relationship values ​​between other nodes in the community set and the preset node, and determining the risk coefficient of each node in the community set based on the relationship values ​​corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset node.

[0046] Step S105: when the target risk coefficient of the target node is greater than a preset risk threshold, it is determined that the target customer account associated with the target risk coefficient is at risk.

[0047] Through the above steps, account data of N customer accounts within a preset time period can be collected to obtain an account data set. Based on the account data set, the community set to which each node belongs is determined, and the nodes are divided into the community sets to which they belong to obtain M community sets, and the preset nodes in each community set are determined. Based on the account data set, the relationship values ​​between other nodes in the community set and the preset nodes are determined, and the risk coefficient of each node in the community set is determined based on the relationship values ​​corresponding to the other nodes. When the target risk coefficient of the target node is greater than the preset risk threshold, it is determined that the target customer account associated with the target risk coefficient is at risk. In an embodiment of the present invention, nodes representing customer accounts can be allocated to corresponding community sets based on the collected account data set, and then preset nodes in each community set are determined. Then, based on the account data set, relationship values ​​between other nodes in each community set and preset nodes are determined, and then the risk coefficient of each node in the community set is determined based on the relationship values ​​corresponding to other nodes. If there is a node whose risk coefficient is greater than a preset risk threshold, it is determined that the corresponding customer account is at risk, and the risk is handled in a timely manner. By deeply mining the relationship between customer accounts, the accuracy of risk assessment of customer accounts can be effectively improved, providing a more comprehensive reference basis for risk control and management of financial institutions, thereby solving the technical problem that in the related technology, effective risk assessment of customer accounts of financial institutions cannot be performed, which easily leads to risks for financial institutions.

[0048] The embodiment of the present invention is described in detail below in combination with the above steps.

[0049] Step S101, collecting account data of N customer accounts within a preset time period to obtain an account data set, wherein N is a positive integer and each customer account is represented as a node.

[0050] In an embodiment of the present invention, a large amount of customer data in a financial institution system within a preset time period (such as within the past year) can be first collected (that is, account data of N (N is a positive integer) customer accounts within a preset time period are collected to obtain an account data set), and the collected data can be integrated and cleaned to ensure that the data used in the subsequent construction of the association graph is accurate, complete and reliable.

[0051] In an embodiment of the present invention, the collected customer data (i.e., account data) includes: basic information (i.e., customer attribute information, such as name, age, kinship, etc.), transaction behavior, risk attributes, and external labels and other features. Afterwards, multi-dimensional data such as basic relationship organization, derived relationship reasoning, knowledge graph generation, and graph feature precipitation can be formed based on the customer data.

[0052] In the embodiment of the present invention, since the association graph data involves sensitive information of customers, strict data privacy protection measures need to be taken during the data collection, integration and processing to ensure that customer privacy is effectively protected.

[0053] In the embodiment of the present invention, each customer account may be characterized as a node to facilitate subsequent construction of an association graph.

[0054] Step S102: based on the account data set, determine the community set to which each node belongs, and divide the nodes into the community sets to which they belong, to obtain M community sets, where M is a positive integer.

[0055] Optionally, the step of determining the community set to which each node belongs based on the account data set includes: characterizing each node as an initial community set; traversing all nodes, and determining a first modularity increment value of the first adjacent community set after placing the traversed first node into the first adjacent community set, wherein the first adjacent community set is a community set in which the node distance between the first node and all first adjacent nodes in the first adjacent community set is less than a preset distance threshold, and the first adjacent node is a node in the first adjacent community set; when the first modularity increment value is greater than the preset increment threshold, determining to move the first node into the first adjacent community set, and generating multiple preset community sets when all nodes are traversed; based on the account data set, updating each preset community set until the modularity of all preset community sets no longer increases, thereby obtaining multiple target community sets; and determining the community set to which each node belongs based on the multiple target community sets.

[0056] In an embodiment of the present invention, the association relationship between customers can be modeled through social network analysis technology (that is, based on the account data set, the community set to which each node belongs can be determined, and the nodes can be divided into the community sets to which they belong, so as to obtain M (M is a positive integer) community sets), including: transactions, communications and other relationships between customers, as well as information such as basic attributes and behavioral characteristics of customers.

[0057] In an embodiment of the present invention, a community detection algorithm (e.g., Louvain algorithm) may be used to perform community detection to divide nodes into different communities (a community may be understood as a collection of nodes having the same characteristics). In this way, the structure and relationships in the network may be better understood and discovered.

[0058] In the embodiment of the present invention, the Louvain algorithm (i.e., the Louvain algorithm) is a community detection algorithm based on modularity optimization, which is widely used in social networks, biological networks, transportation networks and other fields. The Louvain algorithm allocates nodes to different communities by iteratively optimizing the modularity value of the network. The specific implementation process can be divided into two stages, namely, initial partitioning and iterative optimization, wherein:

[0059] The initial division stage is: each node forms a community separately (that is, each node is characterized as an initial community set), then traverse all nodes, try to put each node into the community adjacent to it, and calculate the modularity increment value after being put in (that is, traverse all nodes, and determine the first modularity increment value of the first adjacent community set after the traversed first node is put into the first adjacent community set, the first adjacent community set is a community set where the node distance between the first node and all the first adjacent nodes in the first adjacent community set is less than a preset distance threshold (which can be set according to actual conditions), and the first adjacent node is a node in the first adjacent community set). If the modularity increases after being put in, the node is moved to the new community, otherwise no operation is performed (that is, when the first modularity increment value is greater than the preset increment threshold, it is determined to move the first node into the first adjacent community set). After traversing all nodes, all nodes are recombined according to the new community (that is, when all nodes are traversed, multiple preset community sets are generated), and the next stage of iterative optimization process is entered. The iterative optimization process is: according to the account data set, each preset community set is updated until the modularity of all preset community sets no longer increases, so as to obtain multiple target community sets. After that, the community set to which each node belongs can be determined based on the multiple target community sets.

[0060] Optionally, the step of determining a first modularity increment value of the first neighboring community set after the traversed first node is placed into the first neighboring community set includes: determining all first neighboring nodes included in the first neighboring community set before the traversed first node is placed into the first neighboring community set; calculating the first modularity of the first neighboring community set based on account data associated with all first neighboring nodes; after the traversed first node is placed into the first neighboring community set, calculating the second modularity of the first neighboring community set based on account data associated with all first neighboring nodes and account data associated with the first node; and calculating the difference between the second modularity and the first modularity to obtain the first modularity increment value.

[0061] In an embodiment of the present invention, before placing the traversed first node into the first neighboring community set, all first neighboring nodes included in the first neighboring community set may be determined first, and then the first modularity of the first neighboring community set is calculated based on the account data associated with all the first neighboring nodes. Then, after placing the traversed first node into the first neighboring community set, the second modularity of the first neighboring community set needs to be calculated based on the account data associated with all the first neighboring nodes and the account data associated with the first node. Then, the difference between the second modularity and the first modularity is calculated, and the difference is the first modularity increment value.

[0062] In the embodiment of the present invention, the modularity of a community set is used to indicate the association relationship between nodes in the community set. The closer the nodes are, the higher the modularity.

[0063] Optionally, based on the account data set, each preset community set is updated until the modularity of all preset community sets no longer increases, and the step of obtaining multiple target community sets includes: traversing the nodes in all preset community sets, and determining a second modularity increment value of the second adjacent community set after the traversed second node is placed in the second adjacent community set, wherein the second adjacent community set is a preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than a preset distance threshold, and the second adjacent node is a node in the second adjacent community set; when the second modularity increment value is greater than the preset increment threshold, determining to move the second node into the second adjacent community set; and continuing to determine a third modularity increment value of the third adjacent community set after the traversed third node is placed in the third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, and multiple target community sets are obtained.

[0064] In the embodiment of the present invention, the iterative optimization stage is: according to the current community division, the modularity of each community is calculated, and all nodes are traversed, each node is tried to be placed in the community adjacent to it, and the modularity increment value after being placed is calculated (that is, the nodes in all preset community sets are traversed, and the second modularity increment value of the second adjacent community set is determined after the traversed second node is placed in the second adjacent community set, and the second adjacent community set is a preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than the preset distance threshold, and the second adjacent node is a node in the second adjacent community set). If the modularity increases after being placed, the node is moved to the new community, otherwise no operation is performed (that is, when the second modularity increment value is greater than the preset increment threshold, it is determined to move the second node to the second adjacent community set). The process will be iterated continuously until the community no longer changes or the modularity no longer increases, and the final output result is a community division (that is, continue to determine the third modularity increment value of the third adjacent community set after the traversed third node is placed in the third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, and multiple target community sets are obtained). Among them, each community is a subgraph, the connection between communities is weak, and the connection within the community is relatively dense.

[0065] Step S103: determining the preset nodes in each community set.

[0066] Optionally, the step of determining the preset nodes in each community set includes: determining a first center value of each node in the community set, and selecting nodes indicated by the first center value greater than a first preset center threshold to obtain a first node set; determining a second center value of each node in the community set, and selecting nodes indicated by the second center value greater than a second preset center threshold to obtain a second node set; determining a third center value of each node in the community set, and selecting nodes indicated by the third center value greater than a third preset center threshold to obtain a third node set; performing an intersection operation on the first node set, the second node set, and the third node set to obtain a fourth node set; when the fourth node set includes multiple nodes, selecting a node indicated by the maximum average center value as the preset node in the community set, wherein the average center value is an average value among the first center value, the second center value, and the third center value.

[0067] In the embodiment of the present invention, the centrality measurement algorithm is an algorithm for identifying the most important nodes in a community network. Centrality measurement is an important method in social network analysis, which is used to evaluate the importance and influence of nodes in a community network. Centrality measurement algorithms include: degree centrality, closeness centrality, betweenness centrality, etc.

[0068] In an embodiment of the present invention, a centrality measurement algorithm can be used to determine the preset nodes (i.e., the most important nodes) in each community set. Specifically, the first centrality value (i.e., the degree centrality value) of each node in the community set can be first determined, and the nodes indicated by the first centrality value greater than a first preset centrality threshold (which can be set according to specific circumstances) are selected to obtain a first node set (i.e., some nodes with relatively high degree centrality values ​​are screened out), and the second centrality value (i.e., the proximity centrality value) of each node in the community set can be determined, and the nodes indicated by the second centrality value greater than a second preset centrality threshold (which can be set according to specific circumstances) are selected to obtain a second node set (i.e., some nodes with relatively high proximity centrality values ​​are screened out), and at the same time, the third centrality value (i.e., the betweenness centrality value) of each node in the community set can be determined, and the nodes indicated by the third centrality value greater than a third preset centrality threshold (which can be set according to specific circumstances) are selected to obtain a third node set (i.e., some nodes with relatively high betweenness centrality values ​​are screened out). The first node set, the second node set and the third node set are subjected to an intersection operation to obtain a fourth node set (i.e., some nodes with relatively high degree centrality, closeness centrality and betweenness centrality are screened out). If the fourth node set includes multiple nodes, the node indicated by the maximum average centrality value can be selected as the preset node in the community set (i.e., the node indicated by the maximum average value among degree centrality, closeness centrality and betweenness centrality is finally selected as the most important node), and the average centrality value is the average value among the first centrality value, the second centrality value and the third centrality value.

[0069] In an embodiment of the present invention, the centrality measurement algorithm can be used to help identify key nodes in a community network and evaluate the importance and influence of nodes in the network.

[0070] Optionally, the risk assessment method also includes: determining the number of nodes in the community set and the node degrees of the nodes in the community set, and determining a first central value of the node based on the number of nodes and the node degrees; determining the shortest path distance between the node in the community set and the remaining nodes, and determining a second central value of the node based on the number of nodes and all shortest path distances, wherein the remaining nodes are nodes in the community set other than the node; determining the number of paths through which the shortest path between any two nodes in the community set passes, and determining a third central value of the node based on the number of nodes and the number of paths.

[0071] In the embodiment of the present invention, degree centrality refers to the number of connections that a node has with other nodes in the network. A node with high degree centrality usually means that the node has more extensive connections and influence in the network. Degree centrality DCi can be calculated using the following formula (i.e., first determine the number of nodes in the community set and the node degrees of the nodes in the community set, and then determine the first central value of the node based on the number of nodes and the node degrees):

[0072]

[0073] Among them, Ki represents the number of edges connected to node i (that is, the node degree of the nodes in the community set); N represents the total number of nodes in the community network to which node i belongs (that is, the number of nodes in the community set).

[0074] Closeness centrality refers to the average distance from a node to other nodes. A node with high closeness centrality usually means that the node can communicate and transmit information with other nodes faster. Closeness centrality CCi can be calculated using the following formula (i.e., the shortest path distance between the node in the community set and the remaining nodes (nodes in the community set other than the node) can be determined first, and then the second central value of the node can be determined based on the number of nodes and all shortest path distances):

[0075]

[0076]

[0077] Among them, dij represents the shortest path distance between node i and other nodes j, di represents the average distance from node i to the other points, and the reciprocal of the average distance CCi is the closeness center CCi.

[0078] Betweenness centrality refers to the number of shortest paths that a node has to connect to other nodes in the network. A node with high betweenness centrality usually means that the node plays an important role as a bridge and connection in the network. i It can be calculated using the following formula (i.e., the number of paths through which the shortest path between any two nodes in the community set passes through the nodes can be determined first, and the third center value of the node can be determined based on the number of nodes and the number of paths):

[0079]

[0080] in, represents the number of paths through which the shortest path between node s and node t passes through node i (i.e., the number of paths through which the shortest path between any two nodes in the community set passes through the node); g st Represents the number of shortest paths connecting node s and node t.

[0081] Step S104, based on the account data set, determining the relationship values ​​between other nodes in the community set and the preset node, and determining the risk coefficient of each node in the community set based on the relationship values ​​corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset node.

[0082] In an embodiment of the present invention, a reasonable risk assessment model can be designed based on the association graph data (i.e., the constructed community set), and the risk coefficient can be determined by analyzing the relationship and behavior between customers, thereby realizing accurate risk customer identification and management (i.e., the relationship value between other nodes in the community set (nodes other than the preset nodes in the community set) and the preset nodes can be determined based on the account data set, and the risk coefficient of each node in the community set can be determined based on the relationship values ​​corresponding to the other nodes). That is, according to the association relationship between the nodes in the constructed community set and the determined most important node, the relationship between other nodes and the most important node can be analyzed. If the relationship is close, its own risk coefficient needs to be accumulated.

[0083] In the embodiment of the present invention, the results of the risk assessment model have an important impact on the decision-making of the financial institution, so it is necessary to ensure the interpretability of the model, that is, the financial institution personnel can understand the decision-making process and results of the model, so as to better perform customer management and risk control.

[0084] Optionally, the step of determining the relationship value between other nodes in the community set and the preset node based on the account data set includes: obtaining transaction data between other nodes and the preset node from the account data set; and determining the relationship value based on the transaction data.

[0085] In an embodiment of the present invention, the process of analyzing the closeness of the relationship between other nodes and the most important node is as follows (i.e., determining the relationship value between other nodes in the community set and the preset node based on the account data set): the transaction data between other nodes and the preset node can be first obtained from the account data set, and then the relationship value can be determined based on the transaction data (for example, the larger the transaction amount each time, the higher the transaction frequency, the closer the relationship between the nodes conducting the transaction, and the higher the relationship value).

[0086] Optionally, the step of determining the risk coefficient of each node in the community set based on the relationship values ​​corresponding to other nodes includes: determining a first coefficient for each node in the community set; screening relationship values ​​greater than a preset relationship threshold to obtain a target relationship value set, and determining a second coefficient corresponding to each target relationship value in the target relationship value set; when the node has a corresponding second coefficient, accumulating the first coefficient and the second coefficient to obtain the risk coefficient of the node; or, when the node does not have a second coefficient, characterizing the first coefficient as the risk coefficient of the node.

[0087] In the embodiment of the present invention, each node has its own risk coefficient (i.e., the first coefficient) (i.e., first determine the first coefficient of each node in the community set), and after obtaining the relationship value between each node and the preset node, determine the node with a higher relationship value (i.e., filter the relationship value greater than the preset relationship threshold to obtain the target relationship value set), and then accumulate the risk coefficient of the node according to the relationship value of the node (i.e., determine the second coefficient corresponding to each target relationship value in the target relationship value set, and when the node has a corresponding second coefficient, accumulate the first coefficient and the second coefficient to obtain the risk coefficient of the node). If the relationship value of the node is low, the risk coefficient of the node itself is characterized as the risk coefficient of the node.

[0088] Step S105: when the target risk coefficient of the target node is greater than a preset risk threshold, it is determined that the target customer account associated with the target risk coefficient is at risk.

[0089] In an embodiment of the present invention, if the target risk coefficient of a target node is greater than a preset risk threshold (which can be set according to actual conditions), it can be determined that there is a risk in the target customer account associated with the target risk coefficient, and relevant personnel can be notified in a timely manner to conduct risk investigation and processing.

[0090] Optionally, after collecting account data of N customer accounts within a preset time period and obtaining an account data set, the method further includes: determining a preset number of clusters, and selecting K cluster centers from all nodes, wherein K is a positive integer, K is equal to the preset number of clusters, and a cluster center corresponds to a cluster set; determining the cluster set to which each node belongs, wherein the step of determining the cluster set to which each node belongs includes: determining the distance between the node and the cluster center, and adding the node to the cluster set where the cluster center indicated by the minimum distance is located; when determining the cluster set to which all nodes belong, updating the cluster centers in the cluster set, and executing the step of determining the cluster set to which each node belongs, until the cluster center in each cluster set no longer changes or the number of executions is reached.

[0091] In the embodiment of the present invention, another unsupervised learning community detection algorithm may be used to divide nodes as a supplement to the supervised learning community detection algorithm.

[0092] In an embodiment of the present invention, a group discovery algorithm can identify potential subgroups in a network, thereby better understanding the group structure and relationships in the network. A group discovery algorithm is an algorithm designed to discover clustering structures in a network. It is an unsupervised learning algorithm that can discover and divide data points with similar characteristics (i.e., nodes representing customer accounts) from data without knowing any category information in advance. For example, the K-means clustering algorithm is used to achieve the division of data points, and its implementation process is as follows:

[0093] (1) First, randomly select K initial cluster centers (i.e., determine the preset number of clusters and select K cluster centers from all nodes, where K is a positive integer, K is equal to the preset number of clusters, and a cluster center corresponds to a cluster set).

[0094] (2) Calculate the distance between each data point and the K cluster centers, and assign it to the cluster with the closest cluster center (i.e., determine the cluster set to which each node belongs, wherein the step of determining the cluster set to which each node belongs includes: determining the distance between the node and the cluster center, and adding the node to the cluster set with the cluster center indicated by the minimum distance).

[0095] (3) For each cluster, recalculate the average value of all data points in it and use it as the new cluster center (that is, after determining the cluster set to which all nodes belong, update the cluster center in the cluster set).

[0096] (4) Repeat steps (2) and (3) until the cluster center no longer changes or the preset number of iterations is reached (i.e., the step of determining the cluster set to which each node belongs is executed until the cluster center in each cluster set no longer changes or the number of executions is reached).

[0097] Ultimately, K-means clustering divides all data points into K clusters and assigns each data point to one of the clusters, making the distance between data points in the same cluster as small as possible and the distance between different clusters as large as possible, resulting in different groups.

[0098] Another optional specific implementation is described in detail below.

[0099] This embodiment can use the data of a financial institution as an experiment to extract internal and external data for the past year, including basic information, transaction behavior, risk attributes, and external labels, forming multi-dimensional data such as basic relationship organization, derivative relationship reasoning, knowledge graph generation, and graph feature precipitation. Then use the following algorithm to achieve community division, identify the most important nodes in the network, and identify potential subgroups. The specific process is as follows:

[0100] 1) Use the Louvain algorithm to divide customers into communities.

[0101] (1) Graph construction: The transaction relationship between customers is constructed into an undirected weighted graph. Each customer corresponds to a node in the graph, and the weights between nodes represent transaction frequency or amount.

[0102] (2) Louvain algorithm partitioning: Use the Louvain algorithm to partition the graph into communities and divide customers into different communities. The Louvain algorithm is an algorithm based on modularity optimization that can effectively partition nodes into communities with high cohesion and low coupling.

[0103] (3) Community evaluation: Evaluate the quality of each community by calculating indicators such as modularity to measure the internal tightness and external looseness of the community.

[0104] 2) Use the centrality measurement algorithm to identify the most important nodes in the community network. The steps to use the centrality measurement algorithm to identify the most important nodes in the community network are as follows:

[0105] (1) Graph construction: Build the corresponding graph object according to specific requirements and data types. You can use libraries such as networkx (a Python package that is used to build and manipulate complex graph structures and provides algorithms for analyzing graphs) and igraph (a set of network analysis tools) to build the graph.

[0106] (2) Adding nodes and edges: Add nodes and edges to the graph object based on specific data to build a network structure.

[0107] (3) Calculate centrality: Select an appropriate centrality measurement algorithm, such as betweenness centrality, degree centrality, closeness centrality, etc., and calculate the centrality value of each node.

[0108] (4) Find the most important node: According to the centrality value, find the node with the largest centrality value, which is the most important node.

[0109] (5) Output results: The most important nodes and their corresponding centrality values ​​are output for analysis and visualization.

[0110] 3) Use K-means clustering algorithm to identify subgroups. The specific steps of using K-means clustering algorithm to identify subgroups are as follows:

[0111] (1) Feature standardization: The features are standardized to ensure that the importance of different features is equal.

[0112] (2) Select K value: Select an appropriate K value, that is, the number of clusters, based on the specific task and data type.

[0113] (3) Initialize cluster centers: According to the K value, randomly initialize K cluster centers.

[0114] (4) Calculate the distance: According to the distance measurement method, calculate the distance between each sample point and the cluster center. You can use methods such as Euclidean distance and Manhattan distance.

[0115] (5) Assign samples to clusters: Assign each sample point to the cluster with the nearest cluster center.

[0116] (6) Update cluster centers: Recalculate the coordinates of cluster centers based on the sample points in each cluster.

[0117] (7) Repeat iteration: Repeat steps (4) to (6) until the clustering results converge or the maximum number of iterations is reached.

[0118] Finally, by combining the above three algorithms, it is possible to construct a risk customer association map based on social network analysis technology, and identify the customer’s own risks and potential risks.

[0119] In the embodiment of the present invention, the construction of a risk customer association map based on social network analysis technology can achieve the following beneficial effects: (1) A more comprehensive customer association relationship construction is achieved: the risk assessment model in the related technology usually only considers the risk situation of the customer itself, while the present embodiment adopts social network analysis technology to take the customer's association relationship into consideration, which can more comprehensively assess the risk; (2) A more accurate risk assessment is achieved: a variety of centrality measurement algorithms and group discovery algorithms can be used to conduct more accurate risk assessment and classification of customers, which helps financial institutions to formulate risk control strategies more accurately; (3) More efficient data processing is achieved: the use of a storage method based on a graph database and an efficient community detection algorithm such as the Louvain algorithm can quickly build a customer association map and perform analysis; (4) A better user experience is achieved: by analyzing the customer association relationship, personalized services and recommendations can be provided to customers, improving user experience and loyalty; (5) A better application prospect is achieved: social network analysis technology has broad application prospects in the financial field and can provide financial institutions with better customer relationship management and risk control methods.

[0120] The following is a detailed description in conjunction with another embodiment.

[0121] Embodiment 2

[0122] A risk assessment device for a financial account provided in this embodiment includes multiple implementation units, each of which corresponds to each implementation step in the above-mentioned embodiment 1.

[0123] Figure 2 is a schematic diagram of an optional risk assessment device for a financial account according to an embodiment of the present invention, such as Figure 2 As shown, the risk assessment device may include: a collection unit 20, a first determination unit 21, a second determination unit 22, a third determination unit 23, and a fourth determination unit 24, wherein:

[0124] A collection unit 20 is used to collect account data of N customer accounts within a preset time period to obtain an account data set, where N is a positive integer and each customer account is represented as a node;

[0125] A first determining unit 21 is used to determine the community set to which each node belongs based on the account data set, and divide the nodes into the community sets to which they belong, to obtain M community sets, where M is a positive integer;

[0126] A second determining unit 22, configured to determine a preset node in each community set;

[0127] A third determining unit 23 is used to determine the relationship value between other nodes in the community set and the preset node based on the account data set, and determine the risk coefficient of each node in the community set based on the relationship value corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset node;

[0128] The fourth determining unit 24 is configured to determine that a target customer account associated with the target risk coefficient is at risk when the target risk coefficient of the target node is greater than a preset risk threshold.

[0129] The above-mentioned risk assessment device can collect account data of N customer accounts within a preset time period through the collection unit 20 to obtain an account data set, determine the community set to which each node belongs based on the account data set through the first determination unit 21, and divide the nodes into the community sets to which they belong to obtain M community sets, determine the preset nodes in each community set through the second determination unit 22, determine the relationship values ​​between other nodes in the community set and the preset nodes based on the account data set through the third determination unit 23, and determine the risk coefficient of each node in the community set based on the relationship values ​​corresponding to the other nodes, and determine that the target customer account associated with the target risk coefficient is at risk when the target risk coefficient of the target node is greater than the preset risk threshold through the fourth determination unit 24. In an embodiment of the present invention, nodes representing customer accounts can be allocated to corresponding community sets based on the collected account data set, and then preset nodes in each community set are determined. Then, based on the account data set, relationship values ​​between other nodes in each community set and preset nodes are determined, and then the risk coefficient of each node in the community set is determined based on the relationship values ​​corresponding to other nodes. If there is a node whose risk coefficient is greater than a preset risk threshold, it is determined that the corresponding customer account is at risk, and the risk is handled in a timely manner. By deeply mining the relationship between customer accounts, the accuracy of risk assessment of customer accounts can be effectively improved, providing a more comprehensive reference basis for risk control and management of financial institutions, thereby solving the technical problem that in the related technology, effective risk assessment of customer accounts of financial institutions cannot be performed, which easily leads to risks for financial institutions.

[0130] Optionally, the first determination unit includes: a first characterization module, used to characterize each node as an initial community set; a first traversal module, used to traverse all nodes and determine a first modularity increment value of the first adjacent community set after the traversed first node is placed into the first adjacent community set, wherein the first adjacent community set is a community set in which the node distance between the first node and all first adjacent nodes in the first adjacent community set is less than a preset distance threshold, and the first adjacent node is a node in the first adjacent community set; a first determination module, used to determine to move the first node into the first adjacent community set when the first modularity increment value is greater than a preset increment threshold, and generate multiple preset community sets when all nodes are traversed; a first update module, used to update each preset community set based on the account data set until the modularity of all preset community sets no longer increases, thereby obtaining multiple target community sets; a second determination module, used to determine the community set to which each node belongs based on the multiple target community sets.

[0131] Optionally, the first traversal module includes: a first determination submodule, used to determine all first neighboring nodes included in the first neighboring community set before placing the traversed first node into the first neighboring community set; a first calculation submodule, used to calculate the first modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes; a second calculation submodule, used to calculate the second modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes and the account data associated with the first node after placing the traversed first node into the first neighboring community set; and a third calculation submodule, used to calculate the difference between the second modularity and the first modularity to obtain a first modularity increment value.

[0132] Optionally, the first updating module includes: a first traversal submodule, used to traverse the nodes in all preset community sets, and determine a second modularity increment value of the second adjacent community set after the traversed second node is placed into the second adjacent community set, wherein the second adjacent community set is a preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than a preset distance threshold, and the second adjacent node is a node in the second adjacent community set; a second determination submodule, used to determine to move the second node into the second adjacent community set when the second modularity increment value is greater than the preset increment threshold; and a third determination submodule, used to continue to determine a third modularity increment value of the third adjacent community set after the traversed third node is placed into the third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, thereby obtaining multiple target community sets.

[0133] Optionally, the second determination unit includes: a third determination module, used to determine the first center value of each node in the community set, and select the nodes indicated by the first center value greater than the first preset center threshold to obtain a first node set; a fourth determination module, used to determine the second center value of each node in the community set, and select the nodes indicated by the second center value greater than the second preset center threshold to obtain a second node set; a fifth determination module, used to determine the third center value of each node in the community set, and select the nodes indicated by the third center value greater than the third preset center threshold to obtain a third node set; a first operation module, used to perform an intersection operation on the first node set, the second node set and the third node set to obtain a fourth node set; a first selection module, used to select the node indicated by the maximum average center value as the preset node in the community set when the fourth node set includes multiple nodes, wherein the average center value is the average value between the first center value, the second center value and the third center value.

[0134] Optionally, the risk assessment device also includes: a sixth determination module, used to determine the number of nodes in the community set and the node degrees of the nodes in the community set, and determine the first center value of the node based on the number of nodes and the node degrees; a seventh determination module, used to determine the shortest path distance between the node in the community set and the remaining nodes, and determine the second center value of the node based on the number of nodes and all shortest path distances, wherein the remaining nodes are nodes in the community set other than the node; an eighth determination module, used to determine the number of paths through which the shortest path between any two nodes in the community set passes through the nodes, and determine the third center value of the node based on the number of nodes and the number of paths.

[0135] Optionally, the third determination unit includes: a first acquisition module, used to obtain transaction data between other nodes and the preset node from the account data set; and a ninth determination module, used to determine the relationship value based on the transaction data.

[0136] Optionally, the third determination unit also includes: a tenth determination module, used to determine the first coefficient of each node in the community set; a first screening module, used to screen the relationship values ​​greater than the preset relationship threshold, obtain the target relationship value set, and determine the second coefficient corresponding to each target relationship value in the target relationship value set; a first accumulation module, used to accumulate the first coefficient and the second coefficient to obtain the risk coefficient of the node when the node corresponds to the second coefficient; and a second characterization module, used to characterize the first coefficient as the risk coefficient of the node when the node does not have the second coefficient.

[0137] Optionally, the risk assessment device also includes: an eleventh determination module, which is used to determine the preset number of clusters after obtaining the account data set by collecting account data of N customer accounts within a preset time period, and select K cluster centers from all nodes, wherein K is a positive integer, K is equal to the preset number of clusters, and the cluster center corresponds to a cluster set; a twelfth determination module, which is used to determine the cluster set to which each node belongs, wherein the step of determining the cluster set to which each node belongs includes: determining the distance between the node and the cluster center, and adding the node to the cluster set where the cluster center indicated by the minimum distance is located; a second updating module, which is used to update the cluster centers in the cluster set when determining the cluster set to which all nodes belong, and execute the step of determining the cluster set to which each node belongs until the cluster center in each cluster set no longer changes or the number of executions is reached.

[0138] The above-mentioned risk assessment device may also include a processor and a memory. The above-mentioned acquisition unit 20, the first determination unit 21, the second determination unit 22, the third determination unit 23, the fourth determination unit 24, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0139] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters can be adjusted to determine that the target customer account associated with the target risk coefficient is at risk when the target risk coefficient of the target node is greater than a preset risk threshold.

[0140] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0141] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps: collecting account data of N customer accounts within a preset time period to obtain an account data set, determining the community set to which each node belongs based on the account data set, and dividing the nodes into the community sets to which they belong to obtain M community sets, and determining the preset nodes in each community set; determining the relationship value between other nodes in the community set and the preset nodes based on the account data set, and determining the risk coefficient of each node in the community set based on the relationship values ​​corresponding to the other nodes, and when the target risk coefficient of the target node is greater than the preset risk threshold, determining that the target customer account associated with the target risk coefficient is at risk.

[0142] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned financial account risk assessment method.

[0143] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the above-mentioned risk assessment method for a financial account.

[0144] Figure 3 1 is a hardware structure block diagram of an electronic device (or mobile device) for a risk assessment method for a financial account according to an embodiment of the present invention. Figure 3 As shown, the electronic device may include one or more ( Figure 3 302a, 302b, ..., 302n are used to illustrate) processor 302 (processor 302 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), memory 304 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a keyboard, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 3 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 3 More or fewer components as shown, or with Figure 3 Different configurations shown.

[0145] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0146] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0148] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0149] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0150] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0151] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A risk assessment method for a financial account, characterized in that: include: Collect account data of N customer accounts within a preset time period to obtain an account data set, where N is a positive integer and each of the customer accounts is represented as a node; Based on the account data set, determine the community set to which each of the nodes belongs, and divide the nodes into the community sets to which they belong, to obtain M community sets, where M is a positive integer; Determining a preset node in each of the community sets; Based on the account data set, determining a relationship value between other nodes in the community set and the preset node, and determining a risk coefficient of each of the nodes in the community set based on the relationship value corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset node; When the target risk coefficient of the target node is greater than a preset risk threshold, it is determined that the target customer account associated with the target risk coefficient is at risk.

2. The risk assessment method according to claim 1, characterized in that: The step of determining the community set to which each of the nodes belongs based on the account data set includes: Characterizing each of the nodes as an initial community set; Traversing all the nodes, and determining a first modularity increment value of the first neighboring community set after the traversed first node is placed into the first neighboring community set, wherein the first neighboring community set is the community set in which the node distance between the first node and all the first neighboring nodes in the first neighboring community set is less than a preset distance threshold, and the first neighboring node is a node in the first neighboring community set; When the first modularity increment value is greater than a preset increment threshold, determine to move the first node into the first adjacent community set, and generate multiple preset community sets when all the nodes are traversed; Based on the account data set, updating each of the preset community sets until the modularity of all the preset community sets no longer increases, thereby obtaining multiple target community sets; The community set to which each of the nodes belongs is determined based on the multiple target community sets.

3. The risk assessment method according to claim 2, characterized in that: The step of determining a first modularity increment value of the first neighboring community set after the traversed first node is placed into the first neighboring community set comprises: Determine all the first neighboring nodes included in the first neighboring community set before putting the traversed first node into the first neighboring community set; Calculating a first modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes; After placing the traversed first node into the first neighboring community set, calculating a second modularity of the first neighboring community set based on the account data associated with all the first neighboring nodes and the account data associated with the first node; The difference between the second modularity and the first modularity is calculated to obtain an incremental value of the first modularity.

4. The risk assessment method according to claim 2, characterized in that: Based on the account data set, updating each of the preset community sets until the modularity of all the preset community sets no longer increases, and obtaining a plurality of target community sets comprises: Traversing the nodes in all the preset community sets, and determining a second modularity increment value of the second adjacent community set after the traversed second node is placed into the second adjacent community set, wherein the second adjacent community set is the preset community set in which the node distance between the second node and all second adjacent nodes in the second adjacent community set is less than the preset distance threshold, and the second adjacent node is a node in the second adjacent community set; When the second modularity increment value is greater than the preset increment threshold, determining to move the second node into the second neighboring community set; Continue to determine a third modularity increment value of the third adjacent community set after the traversed third node is placed into a third adjacent community set adjacent to the third node, until the modularity of all current community sets no longer increases, thereby obtaining multiple target community sets.

5. The risk assessment method according to claim 1, characterized in that: The step of determining a preset node in each of the community sets includes: Determine a first center value of each of the nodes in the community set, and select the nodes indicated by the first center value greater than a first preset center threshold to obtain a first node set; Determine a second center value of each of the nodes in the community set, and select the nodes indicated by the second center value greater than a second preset center threshold to obtain a second node set; Determine a third center value of each of the nodes in the community set, and select the nodes indicated by the third center value greater than a third preset center threshold to obtain a third node set; Performing an intersection operation on the first node set, the second node set, and the third node set to obtain a fourth node set; In the case where the fourth node set includes a plurality of the nodes, the node indicated by the maximum average center value is selected as the preset node in the community set, wherein the average center value is an average value among the first center value, the second center value and the third center value.

6. The risk assessment method according to claim 5, characterized in that: The risk assessment method further comprises: Determining the number of nodes in the community set and the node degrees of the nodes in the community set, and determining the first center value of the node based on the number of nodes and the node degrees; Determine the shortest path distance between the node and the remaining nodes in the community set, and determine the second center value of the node based on the number of nodes and all the shortest path distances, wherein the remaining nodes are nodes in the community set other than the node; The number of paths in which the shortest path between any two nodes in the community set passes through the node is determined, and the third center value of the node is determined based on the number of nodes and the number of paths.

7. The risk assessment method according to claim 1, characterized in that: The step of determining the relationship value between other nodes in the community set and the preset node based on the account data set includes: Acquire transaction data between the other node and the preset node from the account data set; Based on the transaction data, the relationship value is determined.

8. The risk assessment method according to claim 7, characterized in that: The step of determining the risk coefficient of each of the nodes in the community set based on the relationship values ​​corresponding to the other nodes includes: Determining a first coefficient of each of the nodes in the community set; Screening the relationship values ​​that are greater than a preset relationship threshold to obtain a target relationship value set, and determining a second coefficient corresponding to each target relationship value in the target relationship value set; In the case where the node corresponds to the second coefficient, the first coefficient and the second coefficient are accumulated to obtain the risk coefficient of the node; or In a case where the node does not have the second coefficient, the first coefficient is characterized as the risk coefficient of the node.

9. The risk assessment method according to claim 1, characterized in that: After collecting the account data of N customer accounts within a preset time period to obtain an account data set, the following steps are also included: Determine the preset number of clusters, and select K cluster centers from all the nodes, where K is a positive integer. K is equal to the preset number of clusters, and the cluster center corresponds to one cluster set; Determine the cluster set to which each of the nodes belongs, wherein the step of determining the cluster set to which each of the nodes belongs comprises: determining the distance between the node and the cluster center, and adding the node to the cluster set where the cluster center indicated by the minimum distance is located; When the cluster set to which all the nodes belong is determined, the cluster centers in the cluster set are updated, and the step of determining the cluster set to which each node belongs is performed until the cluster center in each cluster set no longer changes or the number of executions is reached.

10. A risk assessment device for a financial account, characterized in that: include: A collection unit, used to collect account data of N customer accounts within a preset time period to obtain an account data set, wherein N is a positive integer and each of the customer accounts is represented as a node; A first determining unit is configured to determine, based on the account data set, a community set to which each of the nodes belongs, and divide the nodes into the community sets to which they belong, to obtain M community sets, wherein: M is a positive integer; A second determining unit, configured to determine a preset node in each of the community sets; a third determining unit, configured to determine, based on the account data set, a relationship value between other nodes in the community set and the preset node, and determine a risk coefficient of each of the nodes in the community set based on the relationship value corresponding to the other nodes, wherein the other nodes are nodes in the community set other than the preset node; The fourth determining unit is used to determine that a target customer account associated with the target risk coefficient is at risk when the target risk coefficient of the target node is greater than a preset risk threshold.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the risk assessment method for a financial account according to any one of claims 1 to 9.

12. An electronic device, characterized in that: The invention comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the risk assessment method for a financial account as described in any one of claims 1 to 9.

Images