Credit card abnormal risk determination method, device and equipment and readable storage medium

Abstract

The application discloses a credit card abnormal risk determination method, device and equipment and a readable storage medium. The method comprises the following steps: acquiring credit card data of a user; establishing a relationship network graph of the credit card based on behavior data of the credit card and community relationship information; acquiring a first probability that each first node is accessed at a first time and a second probability that a second node is accessed through the first node at a second time; calculating a target probability that each second node in the relationship network graph is accessed at the second time based on the first probability and the second probability; calculating a similarity value of any two credit cards based on the target probability of each second node; clustering the nodes based on the similarity value to obtain a credit card identifier set; and determining a first abnormal credit card when a first credit card identifier in the credit card identifier set is the same as a preset credit card identifier. According to the embodiment of the application, the abnormal credit card can be accurately identified, and the prediction accuracy and timeliness are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of computer application, and particularly relates to a credit card abnormal risk determination method, device and equipment and a readable storage medium. BACKGROUND

[0002] With the development of information technology, credit card abnormal behaviors such as credit card fraud have appeared. From the current discovered credit card fraud, it is found that the credit card fraud gang is organized, and the fraud means is professional and concealed.

[0003] The existing monitoring and identification of abnormal credit cards rely on business personnel or rules set based on existing credit card fraud modes to monitor and identify credit cards, but the accuracy of the existing credit card monitoring and identification is relatively low. SUMMARY

[0004] The present application provides a credit card abnormal risk determination method, device, equipment and computer storage medium, which can solve the problem of low accuracy of credit card monitoring and identification.

[0005] In a first aspect, the present application provides a credit card abnormal risk determination method, which comprises the following steps:

[0006] Obtaining credit card data of a user, the credit card data comprising a credit card identifier, behavior data of the credit card and community relationship information between credit cards;

[0007] Establishing a relationship network graph of the credit card based on the behavior data of the credit card and the community relationship information, wherein a node in the relationship network graph is a credit card identifier, and a connection between nodes is a relationship of the credit cards corresponding to the connection nodes;

[0008] Obtaining a first probability that each first node in the relationship network graph is accessed at a first time and a second probability that a second node is accessed through the first node at a second time; wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node;

[0009] Calculating a target probability of accessing each second node in the relationship network graph at the second time based on the first probability and the second probability of each first node;

[0010] Calculating a similarity value of any two credit cards based on the target probability of each second node in the relationship network graph;

[0011] Clustering the nodes in the relationship network graph based on the similarity value of any two second nodes to obtain a clustered credit card identifier set;

[0012] When the first credit card identifier in the credit card identifier set is the same as the preset credit card identifier, it is determined that the credit card corresponding to the first credit card identifier is the first abnormal credit card.

[0013] In one possible implementation, the method further includes:

[0014] According to the target probability of the credit card corresponding to the first credit card identifier in the credit card identifier set and the target probability of the credit card corresponding to the second credit card identifier, the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier is calculated; the second credit card identifier is any credit card identifier in the credit card identifier set except the first credit card identifier;

[0015] Based on the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier, the first risk probability of credit card abnormality is determined.

[0016] In one possible implementation, the method further includes:

[0017] The number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is obtained;

[0018] The average number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is calculated;

[0019] The ratio of the number of credit cards corresponding to the first credit card identifier in each credit card identifier combination to the average number is calculated;

[0020] Based on the ratio and the first risk probability, the second risk probability of credit card abnormality is determined.

[0021] In one possible implementation, before obtaining the second probability of accessing the second node through the first node at the second time, the method further includes:

[0022] Based on the number of nodes connected to the first node, the second probability is calculated.

[0023] In one possible implementation, based on the first probability and the second probability of each first node, the target probability of each second node in the access relationship network graph at the second time is calculated, including:

[0024] Based on the target probability of each second node, the target probabilities of all second nodes are summed up to obtain a total access probability;

[0025] The difference between the target probability at the second time and the total access probability at the first time is calculated;

[0026] When the difference is less than a predetermined error parameter, the target probability of each second node is output.

[0027] In one possible implementation, the nodes in the relationship network graph are clustered based on the similarity values of any two second nodes, to obtain a clustered credit card identifier set, including:

[0028] In the relationship network graph, any target node is taken as a core for clustering. When the similarity between the nodes in the preset domain of the target node and the target node meets a preset condition, and the number of nodes in the preset domain of the target node meets a preset threshold, a credit card identifier set of the clustering with the target node as the center is obtained.

[0029] In a second aspect, the embodiments of the present application provide a credit card abnormal risk determination apparatus, which comprises:

[0030] The obtaining module is configured to obtain credit card data of a user, the credit card data comprising credit card identifiers, behavior data of credit cards, and community relationship information between credit cards.

[0031] The relationship network graph construction module is configured to establish a relationship network graph of credit cards based on the behavior data of credit cards and the community relationship information, wherein the nodes in the relationship network graph are credit card identifiers, and the connection lines between the nodes are the relationships of the credit cards corresponding to the connection line nodes.

[0032] The obtaining module is further configured to obtain a first probability that each first node in the relationship network graph is accessed at a first time, and a second probability that a second node is accessed through the first node at a second time, wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node.

[0033] The calculation module is configured to calculate a target probability of accessing each second node in the relationship network graph at the second time based on the first probability and the second probability of each first node.

[0034] The calculation module is further configured to calculate a similarity value of any two credit cards based on the target probability of each second node in the relationship network graph.

[0035] The clustering module is configured to cluster the nodes in the relationship network graph based on the similarity values of any two second nodes, to obtain a clustered credit card identifier set.

[0036] The identification module is configured to determine that a credit card corresponding to a first credit card identifier is a first abnormal credit card when the first credit card identifier in the credit card identifier set is the same as a preset credit card identifier.

[0037] In one possible implementation, the apparatus further comprises a determination module.

[0038] The computing module is further configured to calculate a similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier according to a target probability of the credit card corresponding to the first credit card identifier in the credit card identifier set and a target probability of the credit card corresponding to the second credit card identifier; and the second credit card identifier is any credit card identifier in the credit card identifier set except the first credit card identifier.

[0039] The determining module is configured to determine a first risk probability of the credit card anomaly based on the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier.

[0040] In one possible embodiment, the apparatus includes:

[0041] The obtaining module is further configured to obtain a number of the credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets.

[0042] The computing module is further configured to calculate an average number of the credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets.

[0043] The computing module is further configured to calculate a ratio of the number of the credit cards corresponding to the first credit card identifier to the average number in each credit card identifier set.

[0044] The determining module is further configured to determine a second risk probability of the credit card anomaly based on the ratio and the first risk probability.

[0045] In one possible embodiment, the computing module is further configured to calculate the second probability based on a number of nodes connected to the first node.

[0046] In one possible embodiment, the computing module includes:

[0047] The total probability calculating unit is configured to sum up the target probabilities of all the second nodes to obtain a total access probability based on the target probabilities of the second nodes.

[0048] The difference calculating unit is configured to calculate a difference between the total access probability at the second time and the target probability at the first time.

[0049] The error judging unit is configured to output the target probability of each second node when the difference is less than a predetermined error parameter.

[0050] In one possible embodiment, the clustering module is specifically configured to:

[0051] In the relationship network graph, any target node is taken as a core to perform clustering, and when a similarity between a node in a preset domain of the target node and the target node meets a preset condition and a number of the nodes in the preset domain of the target node meets a preset threshold, a credit card identifier set of a cluster with the target node as a center is obtained.

[0052] In a third aspect, the embodiments of the present application provide a credit card abnormal risk determination device, the device comprising:

[0053] a processor and a memory storing computer program instructions;

[0054] The processor executes the computer program instructions to implement the credit card abnormal risk determination method of any one of the above.

[0055] In a fourth aspect, the embodiments of the present application provide a computer storage medium, the computer readable storage medium storing computer program instructions, the computer program instructions being executed by the processor to implement the credit card abnormal risk determination method of any one of the above.

[0056] In a fifth aspect, the embodiments of the present application provide a computer program product, characterized in that the instructions in the computer program product are executed by the processor of an electronic device to enable the electronic device to execute the credit card abnormal risk determination method of any one of the above.

[0057] The embodiments of the present application provide a credit card abnormal risk determination method, a generation device, an electronic device, a storage medium and a computer program product. By obtaining credit card data of a user, a relationship network graph of the credit card is established based on behavior data and community relationship information in the credit card data, and based on a first probability that each first node in the relationship network graph is visited at a first time and a second probability that a second node is accessed through the first node at a second time, a target probability of accessing each second node in the relationship network graph is calculated, the similarity of the credit card corresponding to any two nodes is calculated based on the target probability of each second node, and the identification of the credit card is clustered to obtain a credit card identification set. When the first credit card identification in the credit card identification set is the same as the preset credit card identification, it is determined as the first abnormal credit card. Using the credit card abnormal risk determination method of the embodiments of the present application, large-scale processing can be performed, time consumption is reduced, and prediction accuracy and timeliness are improved. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. For those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0059] Figure 1 is a flowchart of a credit card abnormal risk determination method provided by an embodiment of the present application;

[0060] Figure 2 is a relationship network graph between two credit cards provided by an embodiment of the present application;

[0061] Figure 3 is a flowchart of a method for determining an abnormal credit card risk according to another embodiment of the present application;

[0062] Figure 4 is a flowchart of a method for determining a second risk probability according to yet another embodiment of the present application;

[0063] Figure 5 is a structural diagram of an apparatus for determining an abnormal credit card risk according to still another embodiment of the present application;

[0064] Figure 6 is a structural diagram of an apparatus for determining an abnormal credit card risk according to still another embodiment of the present application. DETAILED DESCRIPTION

[0065] The features and exemplary embodiments of the various aspects of the present application will be described in detail below with reference to the drawings. The following detailed description is merely intended to explain the present application, and is not intended to limit the present application. The present application can be implemented without some of the specific details, which are well known to those skilled in the art. The following description of the embodiments is merely intended to provide a better understanding of the present application through showing examples of the present application.

[0066] It should be noted that the terms such as first and second, etc., are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Also, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device. Without more limitations, the elements defined by the statement "include" do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0067] It should be noted that the acquisition, storage, use and processing of data in the embodiments of the present application comply with the relevant provisions of national laws and regulations.

[0068] The existing monitoring and identification of abnormal credit cards rely on business personnel, or rules set based on existing credit card fraud patterns to monitor and identify credit cards, but the accuracy of the existing credit card monitoring and identification is relatively low.

[0069] To solve the prior art problems, the embodiment of the present application provides a credit card abnormal risk determination method, device, equipment and computer storage medium. First, the credit card abnormal risk determination method provided by the embodiment of the present application is introduced.

[0070] Figure 1 The flowchart of the credit card abnormal risk determination method provided by the embodiment of the present application is shown. As shown in Figure 1 The steps are as follows:

[0071] S110, credit card data of a user is obtained, the credit card data including credit card identification, credit card behavior data and community relationship information between credit cards.

[0072] S120, a relationship network graph of the credit cards is established based on the credit card behavior data and the community relationship information, the nodes in the relationship network graph being the credit card identification, and the connection between the nodes being the relationship of the credit cards corresponding to the connection nodes.

[0073] S130, a first probability that each first node in the relationship network graph is visited at a first time and a second probability that a second node is accessed through the first node at a second time are obtained; wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node.

[0074] S140, a target probability that each second node in the relationship network graph is accessed at the second time is obtained based on the first probability and the second probability of each first node.

[0075] S150, a similarity value of any two credit cards is calculated based on the target probability of each second node in the relationship network graph.

[0076] S160, the nodes in the relationship network graph are clustered based on the similarity value of any two second nodes, and a credit card identification set after clustering is obtained.

[0077] S170, when a first credit card identification in the credit card identification set is the same as a preset credit card identification, a credit card corresponding to the first credit card identification is determined as a first abnormal credit card.

[0078] Thus, the credit card relationship network graph is established based on the credit card behavior data and the community relationship information, the target probability of accessing each second node in the relationship network graph is calculated based on the first probability of being accessed at the first time and the second probability of accessing the second node through the first node at the second time, the similarity of the credit cards corresponding to any two nodes is calculated based on the target probability of each second node, the credit card identifier set is obtained by clustering the nodes in the relationship network graph, and when the first credit card identifier in the credit card identifier set is the same as the preset credit card identifier, the first abnormal credit card is determined. Using the credit card abnormal risk determination method in the embodiment, large-scale processing can be performed, time consumption is reduced, abnormal credit cards can be accurately identified, and prediction accuracy and timeliness are improved.

[0079] The specific implementation of each step is described below.

[0080] In S110, the credit card data of the user is obtained, and the credit card data can include credit card identifiers, credit card behavior data, and community relationship information between credit cards.

[0081] In some embodiments, the credit card identifier can include, but is not limited to, the following information: credit card user ID number, credit card user billing address, credit card user work unit, credit card user work unit address, credit card user mobile phone number, credit card user email, credit card user device ID used when applying for a credit card, IP corresponding to the device used by the credit card user when applying for a credit card, and credit card user residential address.

[0082] The credit card behavior data can be credit card transaction data. The community relationship information between credit cards can include, but is not limited to, credit card customer ID number, credit card customer billing address, credit card customer unit, work unit address, work unit contact number, credit card customer mobile phone number, credit card contact information (such as mobile phone number) left when applying for a credit card, credit card customer email, location (such as latitude and longitude) when the customer applies for a credit card, device ID used by the customer when applying for a credit card, IP corresponding to the device used by the customer when applying for a credit card, and credit card customer residential address.

[0083] In some embodiments, the credit card data can be obtained from a legal database dedicated to storing credit card data.

[0084] In S120, a credit card relationship network graph is established based on the credit card behavior data and the community relationship information. The nodes in the relationship network graph can be credit card identifiers, and the edges can be the relationships between the credit card identifiers.

[0085] In some embodiments, each node in the relationship network graph can correspond to a credit card identification, and each edge can represent a relationship between any two credit card identifications. In one embodiment, the relationship network graph can be represented as G(V, E), where V is a set of nodes of the relationship network graph, each node V represents an entity, and E is a set of edges of the relationship network graph, each edge E represents a relationship between entities.

[0086] In one example, the relationship network graph is as shown in Figure 2 Figure 2 A i , B i , C i , D i , E i , F i , G i , H i , I i , J i , K i , and L i are nodes of the relationship network graph. A i represents a credit card customer ID number, B i represents a credit card customer billing address, C i represents a credit card customer work unit, D i represents a work unit address, E i represents a work unit contact phone number, F i represents a credit card customer mobile phone number, G i represents a credit card customer mobile phone number, H i represents a credit card customer email address, I i represents a credit card customer mobile phone number, J i represents a credit card customer mobile phone number, K i represents a credit card customer mobile phone number, and L i represents a credit card customer residential address.

[0087] An edge between two nodes represents a relationship between the two nodes, i.e., a relationship between two credit cards. The edge between two nodes can be represented as <X, Y>, where X and Y represent the two nodes. As shown in Figure 2 , the relationship between credit cards can be represented as: <A i , C i > represents that the credit card user corresponding to credit card identification A i works at work unit C i , and <A i , F i > represents that the credit card user corresponding to credit card identification A i 's mobile phone number is F​i , <A i , G i > represents the credit card identification A i , Gi, <A i , H i > represents the credit card identification A i , H i , <A i , I i > represents the credit card identification A i , I i , <A i , J i > represents the credit card identification A i , J i , <A i , L i > represents the credit card identification A i , L i , <C i , D i > represents the work unit C i , D i , <C i , E i > represents the work unit C i , E i , <J i , K i > represents the device J i , K i , <A i , B i > represents the credit card identification A i , B i .

[0088] In this way, the credit card data and the connection between the nodes can be better reflected.

[0089] In S130, after the relationship network graph of the credit card is established, a first probability that each first node in the relationship network graph is accessed at a first time and a second probability that a second node is accessed through the first node at a second time are obtained, wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node.

[0090] In some embodiments, the first probability that each first node in the relationship network graph is visited at the first time instance can be set according to the number of all nodes in the relationship network graph, and the first probability that all first nodes in the relationship network graph are visited at the first time instance can be the same. The second node visited through the first node at the second time instance can be connected to each first node in the relationship network graph.

[0091] In one example, there are N nodes in the community relationship network graph, and the probability that the first node is visited at the first time instance is P t＝i (X), where the N nodes are visited at the first time instance with the same probability, and the probability that the first node is visited at the first time instance is P t＝i (X) = 1 / N.

[0092] Based on this, in some embodiments, before obtaining the second probability that the second node is visited through the first node at the second time instance, the above S130 can further include:

[0093] Calculating the second probability based on the number of nodes connected to the first node.

[0094] In one example, the first probability that each first node in the relationship network graph is visited at the first time instance is obtained, and the second probability that the second node is visited through the first node at the second time instance is calculated based on the number of nodes connected to the first node. Specifically, the second probability is calculated by the following formula (1):

[0095] P t＝i+1 (Y|X) = 1 / |Link(Y)| (1)

[0096] Where P t＝i+1 (Y|X) represents the probability of visiting node Y through node X at time t = i + 1, and |Link(X)| represents the number of nodes connected to node X.

[0097] In this way, the second probability that the second node is visited through the first node at the second time instance can be calculated by formula (1).

[0098] In S140, the probability that the first node is at the first time instance and the second node is visited through the first node at the second time instance can be calculated by the product of the first probability and the second probability, and the target probability of visiting the second node at the second time instance can be the product of the sum of the first probabilities of all nodes connected to the second node and the second probability.

[0099] As an example, the probability that the first node is at the first time instance and the second node is visited through the first node at the second time instance is calculated by the following formula (2):

[0100] P t＝i (X)*Pt＝i+1 (Y|X) (2)

[0101] wherein, P t＝i (X) represents the probability of visiting node X at time t = i, P t＝i+1 (Y|X) represents the probability of visiting node Y through node X at time t = i + 1.

[0102] The target probability of visiting the second node at the second time can be the sum of the first probabilities of all nodes connected to the second node and the product of the second probabilities, which is calculated by using the following formula (3):

[0103]

[0104] wherein, P t＝i+1 (Y) represents the probability of visiting node Y, Link(Y) represents the set of nodes connected to node Y, |Link(Y)| represents the number of nodes connected to node Y, and X in Link(Y) represents that X is connected to Y.

[0105] Based on this, in some embodiments, the target probability of each second node in the relationship network graph at the second time is calculated based on the first probability and the second probability of each first node, including:

[0106] S141, based on the target probability of each second node, summing up the target probabilities of all second nodes to obtain a total visiting probability;

[0107] S142, calculating the difference between the target probability at the second time and the total visiting probability at the first time;

[0108] S143, when the difference is less than a predetermined error parameter, outputting the target probability of each second node.

[0109] As an example, the difference between the target probability at the second time and the total visiting probability at the first time is calculated by summing up the target probabilities of all second nodes to obtain a total visiting probability, using the following formula (4):

[0110] ∑ YinG P t＝i+1 -∑ YinG P t＝i (4)

[0111] wherein, Y in G represents all nodes in the relationship network graph.

[0112] When the above difference is less than the predetermined error parameter, the iteration is stopped, and the calculation of the target probability of each second node at the next time is stopped. The error parameter can be adjusted, and the number of iterations can be set.

[0113] When the difference is greater than or equal to a predetermined error parameter, the target probability of each second node at the next time is calculated.

[0114] In this way, the performance can be considered while satisfying the error as small as possible, the iteration time is prevented from being too long, and the iteration number is set, and the iteration calculation is stopped when the specified number of times is reached.

[0115] In S150, according to the community relationship network diagram, each credit card can be mapped into each node according to the credit card certificate number, and the similarity value of any two credit cards is calculated according to the target probability of each second node in the relationship network diagram by using the Euclidean formula.

[0116] As an example, each credit card is vectorized, and the vectors corresponding to credit card i and credit card j are <A i ,B i ,C i ,D i ,E i ,F i ,G i ,H i ,I i ,J i ,K i ,L i > and <A j ,B j ,C j ,D j ,E j ,F j ,G j ,H j ,I j ,J j ,K j ,L j >.

[0117] The similarity value is calculated by using the following formula (5):

[0118]

[0119] Where Distance(i,j) represents the similarity value of any two credit cards, A i ,B i ,C i ,D i ,E i ,F i ,G i ,H i ,I i ,J i ,K i ,L irespectively represent the target probabilities of the credit card i corresponding to the nodes A, B, C, D, E, F, G, H, I, J, K, L in the community relationship network graph. j j j j j j j j j j j j respectively represent the target probabilities of the credit card j corresponding to the nodes A, B, C, D, E, F, G, H, I, J, K, L in the community relationship network graph.

[0120] In S160, based on the similarity values of any two second nodes, the nodes in the above relationship network graph are clustered to obtain a clustered credit card identifier set. The credit card identifier set can have multiple sets.

[0121] Based on this, in order to cluster the nodes in the relationship network graph, in some embodiments, the above S160 specifically includes:

[0122] In the above relationship network graph, any target node is taken as the core for clustering. When the similarity between the nodes in the preset domain of the target node and the target node satisfies the preset condition, and the number of nodes in the preset domain of the target node satisfies the preset threshold, a credit card identifier set centered on the target node is obtained, wherein the credit cards in the same credit card identifier set are highly similar.

[0123] In this way, by the predetermined condition centered on the target node, all nodes in the relationship network graph can be clustered to obtain at least one credit card identifier set.

[0124] In S170, when the first credit card identifier in the credit card identifier set is the same as the preset credit card identifier, it is determined that the credit card corresponding to the first credit card identifier is the first abnormal credit card. The preset credit card identifier can be, but is not limited to, a predetermined credit card user certificate number, a predetermined credit card user billing address, a predetermined credit card user work unit, a predetermined credit card user work unit address, a predetermined credit card user mobile phone number, a predetermined credit card user email, a predetermined credit card user device ID used when applying for a credit card, a predetermined credit card user IP corresponding to the device used when applying for a credit card, and a predetermined credit card user residential address.

[0125] ​​​​​​​​​​​In one example, the preset credit card identifier is a set of abnormal credit card identifiers {A}, and it is determined whether the first credit card identifier in the clustered credit card identifier set is in the set {A}. If the first credit card identifier is consistent with the identifier of any credit card in the set {A}, the first credit card is considered as an abnormal credit card.

[0126] If the first credit card identifier is not consistent with the identifier of any credit card in the set {A}, the first credit card is considered as a non-abnormal credit card.

[0127] In this way, the credit card abnormal risk determination method of the present application can be used for large-scale processing, reducing the time consumption, improving the accuracy and timeliness of determining abnormal credit cards.

[0128] In some embodiments, as shown in FIG. 1, the method can further include S180 and S190. Figure 3

[0129] S180, according to the target probability of the credit card corresponding to the first credit card identifier in the credit card identifier set and the target probability of the credit card corresponding to the second credit card identifier, the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier is calculated; wherein the second credit card identifier is any credit card identifier in the credit card identifier set except the first credit card identifier.

[0130] S190, determining the first risk probability of the credit card abnormality based on the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier.

[0131] In S180, in one example, the credit card identifiers in the identifier set can be divided into two subsets {B} and {C}, wherein the subset {B} represents the set of the first credit card identifiers, which is the set of credit cards consistent with the credit card identifiers in the preset credit card identifier set {A}, that is, the set of credit card identifiers determined as abnormal. The subset {C} is the credit card determined as non-abnormal. In this case, in some embodiments, the abnormal risk of the credit card in {C} can also be determined.

[0132] In some embodiments, the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier can be calculated according to the target probability of the credit card corresponding to the first credit card identifier in the credit card identifier set and the target probability of the credit card corresponding to the second credit card identifier, and the similarity can be calculated by using the cosine distance.

[0133] ​Continuing the previous example, the similarity between the credit card corresponding to each credit card identifier in subset {B} and the credit card corresponding to each credit card identifier in subset {C} is calculated based on the target probability of the credit card corresponding to each credit card identifier in subset {B} and the target probability of the credit card corresponding to each credit card identifier in subset {C}. The target probability of the credit card corresponding to any two credit card identifiers in {B} and the target probability of the credit card corresponding to each credit card identifier in {C} are vectorized, and the credit card c i corresponding to the first credit card identifier is represented by the vector <A i ,B i ,C i ,D i ,E i ,F i ,G i ,H i ,I i ,J i ,K i ,L i >, and the credit card c i corresponding to the second credit card identifier is represented by the vector <A j ,B j ,C j ,D j ,E j ,F j ,G j ,H j ,I j ,J j ,K j ,L j >, where c i is taken from the second credit card identifier set, and b i is taken from the first credit card identifier set. The similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier is calculated using formula (6).

[0134]

[0135] where b i and c i represent different credit cards, A i , B i , C i , D i , E i , F i , G i , H i , I i , J i , K i , and L i represent the credit card ci Target probabilities corresponding to nodes A, B, C, D, E, F, G, H, I, J, K, L, A j , B j , C j , D j , E j , F j , G j , H j , I j , J j , K j , L j respectively represent credit cards b i Target probabilities corresponding to nodes A, B, C, D, E, F, G, H, I, J, K, L.

[0136] The cosine distance between the first risk probability credit card to be determined and the abnormal credit card in the same credit card identifier set is calculated, and the abnormal credit card can be one or more, so the cosine distance obtained also has one or more, and a cosine distance set {cos(c i ,b1), cos(c i ,b2), cos(c i ,b3)…cos(c i ,b i )…cos(c i ,b n )} is obtained. All the obtained cosine distances are calculated by using formula (7) to take the average value:

[0137] avg_cos(c i )=(cos(c i ,b1),+cos(c i ,b2)+cos(c i ,b3)+…+cos(c i ,b i )+…+cos(c i ,b n )) / n (7)

[0138] Wherein, b1, b2, b3…b n Indicates an abnormal credit card.

[0139] The average value is normalized by using formula (8) to convert into the first risk probability:

[0140] P(c i )=(1+avg_cos(c i )) / 2 (8)

[0141] Wherein, avg_cos(c i) represents the average value of the cosine distance, P(c i ) represents the first risk probability of the credit card c i , P(c i ) satisfies 0 =< P(c i ) <=1.

[0142] The first risk probability of all credit cards in any credit card identifier set is calculated.

[0143] In this way, the risk of the credit card used by the user can be predicted according to the identified abnormal credit card, the cosine distance between the credit card to be determined for the first risk probability and the abnormal credit card in the same credit card identifier set is obtained, the average value of the obtained cosine distance is taken, and the first risk probability obtained by normalization can predict the credit card risk probability.

[0144] In some embodiments, after obtaining the first risk probability of the credit card, the method can include: Figure 4

[0145] S191, obtaining the number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets;

[0146] S192, calculating the average number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets;

[0147] S193, calculating the ratio of the number of credit cards corresponding to the first credit card identifier in each credit card identifier combination to the average number;

[0148] S194, determining the second risk probability of the credit card anomaly based on the ratio and the first risk probability.

[0149] In some embodiments, the number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is obtained, the average number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is calculated, the ratio of the number of credit cards corresponding to the first credit card identifier in each credit card identifier combination to the average number is calculated, the first risk probability is adjusted based on the product of the ratio and the first risk probability, and finally the minimum value between 1 and the adjusted first risk probability is taken to obtain the second risk probability.

[0150] As an example, the number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is Count1 Count2 Count3…Count i …Count n , where 1 =< i =< n. The number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is aggregated, and formula (9) is used for calculation:

[0151] ​Sum = Count1 + Count2 + Count3 + … + Count i +…Count n (9)

[0152] The average number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets is calculated using formula (10):

[0153] Avg = Sum / n (10)

[0154] The ratio of the number of credit cards corresponding to the first credit card identifier in each credit card identifier set to the average number is calculated using formula (11):

[0155] R = Count i / Avg (11)

[0156] wherein Avg represents the average number of credit cards corresponding to the first credit card identifier, n represents the total number of all credit card identifier sets, Count i represents the number of credit cards corresponding to the first credit card identifier in the i-th credit card identifier set.

[0157] The first risk probability is adjusted based on the product of the ratio and the first risk probability using formula (12), and the second risk probability is finally obtained by taking the minimum value between 1 and the adjusted first risk probability:

[0158] Final-P(c i ) = min(1, R*P(c i )) (12)

[0159] wherein Final-P(c i ) represents the second risk probability, and P(c i ) represents the first risk probability of the credit card c i .

[0160] Thus, the problem that the risk probability of credit cards in a set containing a large number of abnormal credit cards is higher than that of credit cards in a set containing a small number of abnormal credit cards is solved.

[0161] Based on the credit card abnormal risk determination method provided in the above embodiments, the application also provides a specific implementation mode of a credit card abnormal risk determination device. Please refer to the following embodiments.

[0162] First, referring to Figure 5 , the credit card abnormal risk determination device 500 provided by the embodiments of the application includes the following modules:

[0163] The acquisition module 510 is configured to acquire credit card data of a user, the credit card data including credit card identifiers, behavior data of the credit cards, and community relationship information among the credit cards.

[0164] The relationship network graph construction module 520 is configured to establish a relationship network graph of the credit cards based on the behavior data of the credit cards and the community relationship information, wherein nodes in the relationship network graph are the credit card identifiers, and a connection between the nodes is a relationship of the credit cards corresponding to the connected nodes.

[0165] The acquisition module 510 is further configured to acquire a first probability that each first node in the relationship network graph is accessed at a first time and a second probability that a second node is accessed through the first node at a second time, wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node.

[0166] The calculation module 540 is configured to calculate a target probability that each second node in the relationship network graph is accessed at the second time based on the first probability and the second probability of each first node.

[0167] The calculation module 540 is further configured to calculate a similarity value of any two credit cards based on the target probability of each second node in the relationship network graph.

[0168] The clustering module 560 is configured to cluster the nodes in the relationship network graph based on the similarity value of any two second nodes, to obtain a credit card identifier set after clustering.

[0169] The identification module 570 is configured to determine that a credit card corresponding to a first credit card identifier in the credit card identifier set is a first abnormal credit card when the first credit card identifier is identical to a preset credit card identifier.

[0170] Based on this, in some embodiments, the credit card abnormality risk determination apparatus 500 further includes a determination module.

[0171] The calculation module 540 is further configured to calculate a similarity of a credit card corresponding to the first credit card identifier and a credit card corresponding to a second credit card identifier in the credit card identifier set according to the target probability of the credit card corresponding to the first credit card identifier and the target probability of the credit card corresponding to the second credit card identifier, wherein the second credit card identifier is any credit card identifier in the credit card identifier set except the first credit card identifier.

[0172] The determination module is configured to determine a first risk probability of credit card abnormality based on the similarity of the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier.

[0173] Based on this, in some embodiments, in the credit card abnormality risk determination apparatus 500:

[0174] The obtaining module 510 is further configured to obtain the number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets.

[0175] The calculating module 540 is further configured to calculate the average number of credit cards corresponding to the first credit card identifier in the plurality of credit card identifier sets.

[0176] The calculating module 540 is further configured to calculate the ratio of the number of credit cards corresponding to the first credit card identifier to the average number of credit cards corresponding to the first credit card identifier in each credit card identifier set.

[0177] The determining module is further configured to determine the second risk probability of the credit card anomaly based on the ratio and the first risk probability.

[0178] Based on this, in some embodiments, the calculating module 540 is further configured to calculate the second probability based on the number of nodes connected to the first node.

[0179] Based on this, in some embodiments, the calculating module includes:

[0180] The total probability calculation unit is configured to sum up the target probabilities of all the second nodes to obtain a total access probability based on the target probabilities of each second node.

[0181] The difference calculation unit is configured to calculate the difference between the total access probability at the second time and the target probability at the first time.

[0182] The error judgment unit is configured to output the target probability of each second node when the difference is less than a predetermined error parameter.

[0183] Based on this, in some embodiments, the clustering module 560 is specifically configured to:

[0184] In the relationship network graph, any target node is taken as the core for clustering. When the similarity between the nodes in the preset domain of the target node and the target node meets the preset condition, and the number of nodes in the preset domain of the target node meets the preset threshold, a credit card identifier set of the clustering with the target node as the center is obtained.

[0185] The various modules of the credit card anomaly risk determination apparatus provided by the embodiments of the present application can realize Figure 1 、 Figure 3 and Figure 4 provide the functions of each step of the credit card anomaly risk determination method and achieve its corresponding technical effects. For brevity, further description is omitted here.

[0186] Figure 6 A hardware structure schematic diagram of the credit card anomaly risk determination apparatus provided by the embodiments of the present application is shown.

[0187] The credit card abnormality risk determination device can include a processor 601 and a memory 602 storing computer program instructions.

[0188] In particular, the processor 601 can include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to perform one or more of the embodiments of the present application.

[0189] The memory 602 can include a mass storage that is used to store data or instructions. By way of example, and not limitation, the memory 602 can include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a Universal Serial Bus (USB) drive or a combination of two or more of these. The memory 602 can include removable or non-removable (or fixed) media, where appropriate. The memory 602 can be considered a computer readable medium, where appropriate. The memory 602 can be internal or external to the credit card abnormality risk determination device, where appropriate. In particular embodiments, the memory 602 is non-volatile, solid-state memory.

[0190] The memory can include read-only memory (ROM), random access memory (RAM), magnetic disk storage mediums, optical storage mediums, flash memory devices, electrical, optical, or other physically tangible / moφhological memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software that, when executed (by one or more processors), is operable to perform operations described with reference to the methods according to the aspects of the present disclosure.

[0191] The processor 601 implements any of the credit card abnormality risk determination methods described in the above embodiments by reading and executing computer program instructions stored in the memory 602.

[0192] In one example, the credit card abnormality risk determination device can further include a communication interface 603 and a bus 610. As shown, the processor 601, the memory 602, and the communication interface 603 are connected through the bus 610 and complete communication therebetween. Figure 6

[0193] The communication interface 603 is mainly used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the present application. ​

[0194] Bus 610 includes hardware, software, or both, to couple credit card abnormal risk determination device components to each other. For example, but not limited to, the bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association local (VESA) bus, or another suitable bus or a combination of two or more of these. Where suitable, bus 610 can include one or more buses. Although specific busses are described and illustrated in this embodiment, the present application contemplates any suitable bus or interconnect.

[0195] The credit card abnormal risk determination device can perform the credit card abnormal risk determination method in the embodiments of the present application based on the obtained credit card data of the user, thereby realizing the credit card abnormal risk determination method and device described in combination Figures 1 to 5 with the credit card abnormal risk determination method and device.

[0196] In addition, in combination with the credit card abnormal risk determination method in the above embodiments, the embodiments of the present application can provide a computer storage medium to realize. The computer storage medium has computer program instructions stored thereon; the computer program instructions are executed by a processor to realize any one of the credit card abnormal risk determination methods in the above embodiments.

[0197] The present application also provides a computer program product, instructions in the computer program product are executed by a processor of an electronic device, so that the electronic device executes each process of realizing any one of the credit card abnormal risk determination method embodiments.

[0198] It is to be understood that the application is not limited to particular configurations and processes described herein and shown in the drawings. For simplicity, detailed descriptions of known methods and apparatuses are omitted so as not to obscure the disclosure. In the above-described embodiments, several specific steps are described as examples. However, the method processes of the present application are not limited to the specific steps described and illustrated herein, but can include any number of additional steps or changes to the described steps, or can be performed in any order, or in parallel, depending on the circumstances.

[0199] The functional blocks shown in the structural block diagrams described above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, and the like. When implemented in software, the elements of the present application are program or code segments used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. The "machine-readable medium" can include any medium capable of storing or transmitting information. Examples of the machine-readable medium include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact discs (CD-ROM), optical disks, hard disks, optical fiber media, radio frequency (RF) links, and the like. The code segments can be downloaded via a computer network such as the Internet, an intranet, and the like.

[0200] It is also to be understood that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in an order different from the embodiments, or several steps can be performed simultaneously.

[0201] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0202] The above is merely specific implementation of the present application, and those skilled in the art can clearly understand the specific working process of the system, module and unit described above for the convenience and brevity of description, which can refer to the corresponding process in the foregoing method embodiments, and will not be described herein. It should be understood that the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed in the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A credit card abnormal risk determination method characterized by, The method comprises: obtaining credit card data of a user, the credit card data comprising credit card identification, behavior data of the credit card, and community relationship information between credit cards; establishing a relationship network graph of the credit cards based on the behavior data of the credit cards and the community relationship information, wherein nodes in the relationship network graph are credit card identification, and a connection between nodes is a relationship of credit cards corresponding to the connection; obtaining a first probability that each first node in the relationship network graph is visited at a first time, and a second probability that a second node is accessed through the first node at a second time; wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node; calculating a target probability of accessing each second node in the relationship network graph at the second time based on the first probability and the second probability of each first node; calculating a similarity value of any two credit cards based on the target probability of each second node in the relationship network graph; clustering the nodes in the relationship network graph based on the similarity value of any two second nodes to obtain a clustered credit card identification set; when a first credit card identification in the credit card identification set is the same as a preset credit card identification, determining that a credit card corresponding to the first credit card identification is a first abnormal credit card; the calculation of the target probability of accessing each second node in the relationship network graph at the second time based on the first probability and the second probability of each first node comprises: calculating the product of the sum of the first probabilities of all nodes connected to the second node and the second probability to obtain the target probability of accessing each second node in the relationship network graph at the second time; summing up the target probabilities of all second nodes based on the target probability of each second node to obtain a total access probability; calculating a difference value between the total access probability at the second time and the total access probability at the first time; when the difference value is less than a predetermined error parameter, outputting the target probability of each second node.

2. The credit card anomaly risk determination method of claim 1, wherein, The method further comprises: calculating a similarity between a credit card corresponding to the first credit card identification and a credit card corresponding to a second credit card identification according to the target probability of the credit card corresponding to the first credit card identification and the target probability of the credit card corresponding to the second credit card identification in the credit card identification set; wherein the second credit card identification is any credit card identification in the credit card identification set except the first credit card identification; determining a first risk probability of credit card anomaly based on the similarity between the credit card corresponding to the first credit card identification and the credit card corresponding to the second credit card identification.

3. The credit card anomaly risk determination method of claim 2, wherein, The method further comprises: obtaining a number of credit cards corresponding to a first credit card identification in a plurality of credit card identification sets; calculating an average number of credit cards corresponding to the first credit card identification in the plurality of credit card identification sets; calculating a ratio of the number of credit cards corresponding to the first credit card identification in each credit card identification set to the average number; determining a second risk probability of credit card anomaly based on the ratio and the first risk probability.

4. The credit card anomaly risk determination method of claim 1, wherein, Before obtaining the second probability of accessing to a second node through the first node at a second time, the method further comprises: calculating the second probability based on the number of nodes connected to the first node.

5. The credit card anomalous risk determination method according to any one of claims 1 to 4, characterized in that, Based on the similarity value of any two second nodes, the nodes in the relationship network graph are clustered to obtain a clustered credit card identifier set, comprising: Taking any target node in the relationship network graph as the core for clustering, when the similarity between the nodes in the preset domain of the target node and the target node meets the preset condition, and the number of nodes in the preset domain of the target node meets the preset threshold, a clustered credit card identifier set is obtained with the target node as the center.

6. An apparatus for determining an abnormal risk of a credit card, characterized by, The device comprises: An acquisition module is configured to acquire credit card data of a user, the credit card data comprising credit card identifiers, behavior data of credit cards, and community relationship information between credit cards; A relationship network graph construction module is configured to establish a relationship network graph of credit cards based on the behavior data of the credit cards and the community relationship information, wherein the nodes in the relationship network graph are credit card identifiers, and the connections between the nodes are the relationships of the credit cards corresponding to the connection nodes; The acquisition module is further configured to acquire a first probability of each first node in the relationship network graph being accessed at a first time, and a second probability of accessing to a second node through the first node at a second time; wherein the first node is any node in the relationship network graph, and the second node is any node in the relationship network graph except the first node; A calculation module is configured to calculate a target probability of accessing each second node in the relationship network graph at the second time based on the first probability and the second probability of each first node; The calculation module is further configured to calculate a similarity value of any two credit cards based on the target probability of each second node in the relationship network graph; A clustering module is configured to cluster the nodes in the relationship network graph based on the similarity value of any two second nodes to obtain a clustered credit card identifier set; An identification module is configured to determine a first abnormal credit card when a first credit card identifier in the credit card identifier set is the same as a preset credit card identifier; The calculation module is configured to: calculate the product of the sum of the first probabilities of all nodes connected to the second node and the second probability to obtain a target probability of accessing each second node in the relationship network graph at the second time; sum all the target probabilities of the second nodes based on the target probability of each second node to obtain a total access probability; calculate the difference between the total access probability at the second time and the total access probability at the first time; output the target probability of each second node when the difference is less than a predetermined error parameter.

7. The credit card anomalous risk determination apparatus of claim 6, wherein, The apparatus further comprises a determination module; the calculation module is further configured to calculate the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier according to the target probability of the credit card corresponding to the first credit card identifier and the target probability of the credit card corresponding to the second credit card identifier in the credit card identifier set; and the second credit card identifier is any credit card identifier in the credit card identifier set except the first credit card identifier. The determination module is configured to determine the first risk probability of the credit card anomaly based on the similarity between the credit card corresponding to the first credit card identifier and the credit card corresponding to the second credit card identifier.

8. The credit card anomaly risk determination apparatus according to claim 7, characterized in that, The acquisition module is further configured to acquire the number of credit cards corresponding to the first credit card identifier in a plurality of credit card identifier sets. The calculation module is further configured to calculate the average number of credit cards corresponding to the first credit card identifier in a plurality of credit card identifier sets. The calculation module is further configured to calculate the ratio of the number of credit cards corresponding to the first credit card identifier in each credit card identifier set to the average number. The determination module is further configured to determine the second risk probability of the credit card anomaly based on the ratio and the first risk probability.

9. The credit card anomaly risk determination apparatus according to claim 6, characterized in that, The calculation module is further configured to calculate the second probability based on the number of nodes connected to the first node.

10. The credit card anomalous risk determination apparatus according to any one of claims 6-9, wherein, The clustering module is specifically configured to: cluster any target node in the relationship network graph as the core, and obtain a credit card identifier set of the target node as the center when the similarity between the nodes in the preset domain of the target node and the target node meets the preset condition, and the number of nodes in the preset domain of the target node meets the preset threshold.

11. A credit card abnormality risk determination device characterized by comprising: The device comprises a processor and a memory storing computer program instructions; The processor executes the computer program instructions to implement the credit card anomaly risk determination method according to any one of claims 1-5.

12. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer program instructions, and the computer program instructions are executed by the processor to implement the credit card anomaly risk determination method according to any one of claims 1-5.

13. A computer program product, characterised in that, The instructions in the computer program product are executed by the processor of the electronic device, so that the electronic device can execute the credit card anomaly risk determination method according to any one of claims 1-5.

Images