Transaction identification methods, devices, electronic equipment and storage media

By constructing a directed graph of transactions and utilizing a supervised transaction identification model, the problems of low accuracy and efficiency in the transaction identification process are solved, and efficient and objective screening of abnormal transaction data is achieved.

CN119089353BActive Publication Date: 2026-11-10CHINA CONSTRUCTION BANK +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411150873.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2026-11-10
Estimated Expiration
2044-08-21

AI Technical Summary

Technical Problem

In existing technologies, the accuracy and efficiency of transaction identification are low, and there is subjectivity, making it difficult to identify abnormal transaction data on a large scale.

Method used

By acquiring transaction accounts and transaction transfer relationships from transaction data, a directed graph of transactions is constructed. The directed graph is then identified and cropped, and a supervised transaction identification model is used to filter out abnormal transaction data.

Benefits of technology

It improves the efficiency and accuracy of identifying abnormal transaction data, reduces identification costs and time, and ensures the objectivity and reliability of transaction identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089353B_ABST
    Figure CN119089353B_ABST
Patent Text Reader

Abstract

This disclosure relates to a transaction identification method, apparatus, electronic device, and storage medium, belonging to the field of information processing technology. The method includes: acquiring first transaction data to be identified, and obtaining transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data; constructing a first directed transaction graph based on the transaction accounts and the transaction transfer relationships between every two transaction accounts; identifying the first transaction data, obtaining a first identification result of the first transaction data, and determining normal transaction data in the first transaction data based on the first identification result; trimming the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph; determining second transaction data based on the nodes in the second directed transaction graph, and identifying the second transaction data to filter out abnormal third transaction data from the second transaction data. Therefore, this disclosure improves the accuracy and efficiency of obtaining abnormal transaction data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure relates to the field of information processing, and more particularly to a transaction identification method, apparatus, electronic device, and storage medium. Background Technology

[0002] Identifying abnormal transaction data from massive amounts of transaction data is often difficult. Related technologies include manual monitoring and identification of abnormal and suspicious transactions, as well as identification of abnormal transaction data through expert-defined rules. However, the above methods are often inefficient for manual identification, unable to identify large-scale transaction data, and cannot guarantee the accuracy and timeliness of abnormal transaction data identification. Identification through rules is often subject to a certain degree of subjectivity. Summary of the Invention

[0003] This disclosure provides a transaction identification method, apparatus, electronic device, computer-readable storage medium, and computer program product to at least solve the problems of low accuracy and efficiency in the transaction identification process in related technologies.

[0004] The technical solution disclosed herein is as follows:

[0005] According to a first aspect of the present disclosure, a transaction identification method is provided, the method comprising: acquiring first transaction data to be identified, and acquiring transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data; constructing a first directed transaction graph based on the transaction accounts and the transaction transfer relationships between every two transaction accounts; identifying the first transaction data, acquiring a first identification result of the first transaction data, and determining normal transaction data in the first transaction data based on the first identification result; trimming the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph; determining second transaction data based on the nodes in the second directed transaction graph, and identifying the second transaction data to filter out third transaction data with anomalies from the second transaction data.

[0006] According to a second aspect of the present disclosure, a transaction identification device is provided, the device comprising: a first acquisition module configured to acquire first transaction data to be identified, and to acquire transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data; a second acquisition module configured to construct a first directed transaction graph based on the transaction accounts and the transaction transfer relationships between every two transaction accounts; a determination module configured to identify the first transaction data, acquire a first identification result of the first transaction data, and determine normal transaction data in the first transaction data based on the first identification result; a trimming module configured to trim the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph; and an identification module configured to determine second transaction data based on nodes in the second directed transaction graph, and to identify the second transaction data to filter out third transaction data with anomalies from the second transaction data.

[0007] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the transaction identification method as described in the first aspect above.

[0008] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided that, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, enables the electronic device to perform the transaction identification method as described in the first aspect above.

[0009] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the transaction identification method as described in the first aspect above.

[0010] The technical solution provided by the embodiments of this disclosure brings at least the following beneficial effects: First transaction data to be identified is obtained, and transaction accounts and transaction transfer relationships between every two transaction accounts are obtained from the first transaction data. Based on the transaction accounts and transaction transfer relationships between every two transaction accounts, a first directed transaction graph is constructed. The first transaction data is identified, and a first identification result of the first transaction data is obtained. Based on the first identification result, normal transaction data in the first transaction data is determined. The first directed transaction graph is pruned based on the normal transaction data to obtain a second directed transaction graph. Second transaction data is determined based on the nodes in the second directed transaction graph, and the second transaction data is identified to filter out abnormal third transaction data from the second transaction data. Therefore, this disclosure can identify massive amounts of transaction data, improve the efficiency of identifying abnormal transaction data, ensure the objectivity of obtaining abnormal transaction identification data, save costs and time in the transaction identification process, and improve the accuracy and reliability of obtaining abnormal transaction identification data.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0013] Figure 1 This is a flowchart illustrating the transaction identification method according to the first embodiment of this disclosure.

[0014] Figure 2 This is a schematic diagram of the first directed graph of transactions disclosed in this publication.

[0015] Figure 3 This is a schematic diagram of the process of training a supervised transaction recognition model according to the first embodiment of this disclosure.

[0016] Figure 4 This is a flowchart illustrating the transaction identification method according to the second embodiment of this disclosure.

[0017] Figure 5 This is a schematic diagram of the second directed graph of the disclosed transaction.

[0018] Figure 6 This is a flowchart illustrating the transaction identification method according to the third embodiment of this disclosure.

[0019] Figure 7 This is a flowchart illustrating the transaction identification method according to the fourth embodiment of this disclosure.

[0020] Figure 8This is a flowchart illustrating the transaction identification method according to the fifth embodiment of this disclosure.

[0021] Figure 9 This is a block diagram of a transaction identification device according to a first embodiment of the present disclosure.

[0022] Figure 10 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0024] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0025] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0026] It should be noted that the acquisition, transmission, storage, use, and processing of data in this disclosure comply with the relevant provisions of laws and regulations. In the embodiments of this disclosure, certain software, components, models, and other existing solutions in the industry may be mentioned. These should be considered as exemplary and are intended only to illustrate the feasibility of implementing the technical solution of this disclosure. However, they do not mean that the applicant has used or necessarily used the solution.

[0027] Figure 1 This is a flowchart illustrating the transaction identification method according to the first embodiment of this disclosure.

[0028] like Figure 1 As shown, the transaction identification method of the first embodiment of this disclosure includes the following steps:

[0029] S101, Obtain the first transaction data to be identified, and obtain the transaction account and the transaction transfer relationship between every two transaction accounts from the first transaction data.

[0030] It should be noted that this disclosure does not limit the specific method for obtaining the first transaction data to be identified, and the appropriate method can be selected according to the actual situation.

[0031] Optionally, the original transaction data to be identified can be obtained from the transaction database. In order to improve the efficiency of the transaction identification process, a transaction data filtering strategy can be pre-set for the transaction parameters in the original transaction data. The original transaction data can be filtered based on the transaction data filtering strategy to obtain the first transaction data to be identified.

[0032] For example, regarding the transaction amount in the original transaction data, a transaction amount threshold par_m is preset, and the original transaction data with a transaction amount less than the transaction amount threshold par_m is filtered out. Regarding the transaction time in the original transaction data, a transaction time period [par_d_min, par_d_max] is preset, and the original transaction data that is not within the transaction time period [par_d_min, par_d_max] is filtered out. Regarding the transaction account in the original transaction data, the original transaction data corresponding to the transaction account that is in the transaction account whitelist can be filtered out.

[0033] In this embodiment of the disclosure, after obtaining the first transaction data, the transaction accounts and the transaction transfer relationships between every two transaction accounts can be obtained from the first transaction data.

[0034] S102, construct the first directed graph of transactions based on the transaction accounts and the transaction transfer relationships between every two transaction accounts.

[0035] Here, there is a directed graph G(V,E), where V is the set of nodes in the directed graph and E is the set of edges in the directed graph. The nodes in the directed graph are transaction accounts, and the edges are the transaction transfer relationships between any two transaction accounts. Each edge carries transaction information, including the transaction amount and the transaction time.

[0036] For example, such as Figure 2 As shown, a first directed graph Graph1 is constructed for transaction accounts A1-A6 and the transaction transfer relationships between every two transaction accounts, where node A... i For trading account, side is trading account A i to transaction account A j Initiate a transfer transaction, where (Mi,Ti) represents the transaction amount and time.

[0037] For example, such as Figure 2 As shown, for trading accounts A1 and A2, trading account A1 initiates a transfer transaction to trading account A2, where the transaction amount is M1 and the transaction time is T1.

[0038] S103, identify the first transaction data, obtain the first identification result of the first transaction data, and determine the normal transaction data in the first transaction data based on the first identification result.

[0039] It should be noted that this disclosure identifies the first transaction data, and the specific method for obtaining the first identification result of the first transaction data is not limited, and can be selected according to the actual situation.

[0040] Optionally, the feature vector of the first transaction data can be obtained, and the feature vector can be input into a pre-trained supervised transaction recognition model to obtain the first transaction label prediction value of the first transaction data. Based on the first transaction label prediction value, the first recognition result of the first transaction data can be obtained.

[0041] For example, if the predicted value of the first transaction tag is greater than a preset first threshold, the first transaction data is determined to be normal transaction data; if the predicted value of the first transaction tag is less than or equal to the preset first threshold, the first transaction data is determined to be abnormal transaction data.

[0042] In this embodiment of the disclosure, after obtaining the first identification result, normal transaction data in the first transaction data can be determined based on the first identification result.

[0043] S104. Based on normal transaction data, the first directed graph of transactions is trimmed to obtain the second directed graph of transactions.

[0044] In this embodiment of the disclosure, a first transaction account and a second transaction account that have a transaction transfer relationship in normal transaction data can be obtained. A first node associated with the first transaction account and a second node associated with the second transaction account are determined in the first directed transaction graph. The edges between the first node and the second node are deleted (pruned) to obtain the second directed transaction graph Graph2.

[0045] For example, after obtaining normal transaction data, the first transaction account A with transaction transfer relationships in the normal transaction data... m Second trading account A n In the first directed graph of the first transaction, determine the first node A associated with the first transaction account. m The second node A associated with the second trading account n , the first node A m With the second node A n The edges between the nodes are deleted to obtain the second directed graph Graph2.

[0046] S105. Based on the nodes in the second transaction directed graph, determine the second transaction data and identify the second transaction data to filter out the third transaction data that is abnormal from the second transaction data.

[0047] In this embodiment of the disclosure, after obtaining the second directed graph of transactions, the nodes in the second directed graph can be divided according to the weight of each node in the second directed graph to obtain the node set of the second directed graph. By obtaining the target transaction label prediction value of each node in the node set, the second identification result of the second transaction data is determined according to the target transaction label prediction value. Based on the second identification result, the third transaction data with anomalies is filtered from the second transaction data.

[0048] For example, given the node set {A} in the directed graph of the second transaction, if the node set {A} includes node A... 1 Node A 2 and node A 3 Node A 1 Node A 2 and node A 3 If the predicted values ​​of the target transaction labels are y1, y2, and y3, then the predicted sum of the target transaction labels for the node set {A} is y1, y2, and y3. sum =y1+y2+y3, if the predicted sum of the target transaction labels is y1+y2+y3, then... sum If the score is less than the preset second threshold "bad_score", then all transaction data in the node set is determined to be abnormal, i.e., node A. 1 Node A 2 and node A 3 All transaction data are abnormal. If the target transaction label prediction and value y are all abnormal, sum If the value is greater than or equal to the preset second threshold bad_score, it is determined that all transaction data in the node set is normal, i.e., node A 1 Node A 2 and node A 3 All transaction data are normal.

[0049] In summary, the transaction identification method provided by the embodiments of this disclosure involves acquiring first transaction data to be identified, obtaining transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data, constructing a first directed transaction graph based on the transaction accounts and transaction transfer relationships between every two transaction accounts, identifying the first transaction data, obtaining a first identification result of the first transaction data, determining normal transaction data in the first transaction data based on the first identification result, trimming the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph, determining second transaction data based on the nodes in the second directed transaction graph, and identifying the second transaction data to filter out abnormal third transaction data from the second transaction data. Therefore, this disclosure can identify massive amounts of transaction data, improve the efficiency of identifying abnormal transaction data, ensure the objectivity of obtaining abnormal transaction identification data, save costs and time in the transaction identification process, and improve the accuracy and reliability of obtaining abnormal transaction identification data.

[0050] Figure 3 This is a flowchart illustrating the transaction identification method according to the second embodiment of this disclosure.

[0051] like Figure 3 As shown, the transaction identification method of the second embodiment of this disclosure includes the following steps:

[0052] S301, Obtain the first transaction data to be identified, and obtain the transaction account and the transaction transfer relationship between every two transaction accounts from the first transaction data.

[0053] S302, construct the first directed graph of transactions based on the transaction accounts and the transaction transfer relationships between every two transaction accounts.

[0054] For details regarding steps S301-S302, please refer to the above embodiments; they will not be repeated here.

[0055] As one possible implementation, based on the above implementation, the specific process of identifying the first transaction data and obtaining the first identification result of the first transaction data in step S103 includes steps S303 to S305:

[0056] S303, obtain the feature vector of the first transaction data, input the feature vector into the pre-trained supervised transaction recognition model, and obtain the predicted value of the first transaction label of the first transaction data.

[0057] As one possible way to achieve this, such as Figure 4 As shown, the specific process for training the supervised transaction recognition model includes steps S401 to S405:

[0058] S401, Obtain the training sample set, wherein each training sample in the training sample set includes sample transaction data and the true value of sample transaction label.

[0059] Specifically, regarding the true value of the sample transaction label, if a transfer transaction transi between any two trading accounts is an abnormal transaction, then the true value y of the sample transaction label is... i If every transfer transaction transi between two trading accounts is a normal transaction, then the true value y of the sample transaction label is 1. i It is 0.

[0060] It should be noted that this disclosure does not limit the specific method for obtaining the training sample set, and the appropriate method can be selected according to the actual situation.

[0061] Optionally, a historical transaction dataset can be obtained, an initial training sample set can be constructed based on the historical transaction dataset, and the initial training sample set can be sampled to obtain a training sample set.

[0062] For example, if a sample account has a history of abnormal transactions, that type of sample account is identified as an abnormal account, and all abnormal accounts are retained. These abnormal accounts are then aggregated to form set {B}. If a sample account has no history of abnormal transactions, that type of account is identified as a normal account. Normal accounts can be extracted according to a ratio of "abnormal account: normal account = 1:r1", where r1 is an adjustable parameter. The extracted normal accounts are then aggregated to form set {C}. Sets {B} and {C} are merged to form set {D} = {B}U{C}. The transaction data of the trading accounts in set {D} is then obtained. All abnormal transaction data in the dataset are aggregated to obtain set {E}. Other transaction data (normal transaction data) in set {D} are aggregated to obtain set {F}. All abnormal transaction data are selected from set {E}. Normal accounts are extracted from set {F} according to the ratio of "abnormal transaction: normal transaction = 1:r2", where r2 is an adjustable parameter. The extracted normal accounts are aggregated to obtain set {G}. Set {E} and set {G} are combined to obtain set {H} = {E}U{G}. The obtained set {H} can be divided into training sample set and test sample set according to a preset ratio.

[0063] S402, obtain the sample feature vector of the sample transaction data, input the sample feature vector into the initial transaction recognition model, and obtain the transaction label prediction value.

[0064] In this embodiment of the disclosure, after obtaining sample transaction data, feature extraction can be performed on the sample transaction data to obtain the sample feature vector of the sample transaction data.

[0065] It should be noted that after obtaining the sample transaction data, the true values ​​of the sample transaction labels, and the sample feature vectors of the sample transaction data, a wide table can be constructed based on the sample transaction data, the true values ​​of the sample transaction labels, and the sample feature vectors of the sample transaction data.

[0066] For example, as shown in Table 1, transi represents a transfer transaction between every two trading accounts, FeaVarj is the feature variable corresponding to transi, and x i,j Let FeaVarj be the feature value of the feature variable corresponding to transi, and yi be the true value of the sample transaction label.

[0067] Table 1

[0068]

[0069] In this embodiment of the disclosure, after obtaining the sample feature vector of the sample transaction data, the sample feature vector can be input into the initial transaction recognition model to obtain the transaction label prediction value.

[0070] S403. Based on the actual values ​​of the sample transaction labels and the predicted values ​​of the transaction labels, the initial transaction recognition model is trained in a supervised manner to obtain a supervised transaction recognition model.

[0071] In this embodiment of the disclosure, the target loss of the initial transaction recognition model can be determined based on the true value and predicted value of the sample transaction label. The model parameters of the initial transaction recognition model are adjusted based on the target loss. The adjusted initial transaction recognition model is then trained in a supervised manner based on the next training sample until the training is completed and a supervised transaction recognition model is obtained.

[0072] It should be noted that this disclosure does not limit the specific methods of supervised training. For example, supervised training methods include, but are not limited to, logistic regression algorithm, random forest algorithm, and decision tree-based ensemble machine learning algorithm (eXtremegradient boosting, abbreviated as Xgboost).

[0073] It should be noted that this disclosure does not impose any restrictions on the setting of the model training termination condition, and the model training termination condition can be set according to the actual situation.

[0074] Optionally, the model training termination condition can be set to the target loss value being less than a preset loss threshold; alternatively, the model training termination condition can also be set to the number of times the model parameters of the initial transaction recognition model are adjusted reaches a preset number threshold.

[0075] In this embodiment of the disclosure, the feature vector of the first transaction data can be input into a pre-trained supervised transaction recognition model to obtain the predicted value y of the first transaction label of the first transaction data.m .

[0076] Among them, when the predicted value of the first transaction label is y m The larger the value, the lower the probability that the first transaction data is abnormal.

[0077] S304, in response to the first transaction tag prediction value being greater than a preset first threshold, the first transaction data is determined to be normal transaction data.

[0078] It should be noted that this disclosure does not impose any restrictions on the setting of the first threshold par_s, which can be set according to the actual situation.

[0079] In this embodiment of the disclosure, when the predicted value of the first transaction tag is y m When y is greater than the preset first threshold par_s, i.e. m When > par_s, the first transaction data is determined to be normal transaction data.

[0080] S305, in response to the first transaction tag prediction value being less than or equal to a preset first threshold, the first transaction data is determined to be abnormal transaction data.

[0081] In this embodiment of the disclosure, when the predicted value of the first transaction tag is y m When y is less than or equal to the preset first threshold par_s, i.e. m When the value is less than or equal to par_s, the first transaction data is determined to be abnormal normal transaction data.

[0082] S306, Based on the first identification result, determine the normal transaction data in the first transaction data.

[0083] As one possible implementation, based on the above implementation, the specific process of trimming the first directed graph based on normal transaction data to obtain the second directed graph in step S104 includes steps S307 to S308:

[0084] S307, retrieve the first and second transaction accounts that have a transaction transfer relationship in the normal transaction data.

[0085] S308, in the first directed graph of transactions, determine the first node associated with the first transaction account and the second node associated with the second transaction account, and delete the edge between the first node and the second node to obtain the second directed graph of transactions.

[0086] For example, such as Figure 5As shown, after obtaining normal transaction data, the first transaction account A5 and the second transaction account A6 that have transaction transfer relationships in the normal transaction data are identified. In the first directed graph of transactions, the first node A5 associated with the first transaction account and the second node A6 associated with the second transaction account are determined. The edge between the first node and the second node is deleted, that is, the edge between the first node A5 and the second node A6 is deleted, so as to obtain the second directed graph of transactions Graph2.

[0087] S309, Determine the second transaction data based on the nodes in the directed graph of the second transaction.

[0088] For details regarding step S309, please refer to the above embodiments; they will not be repeated here.

[0089] As one possible implementation, based on the above implementation, the specific process of identifying the second transaction data in step S105 to filter out the third transaction data that is abnormal includes steps S3010 to S3013:

[0090] S3010: Obtain the transaction probability of each node in the second transaction directed graph, and use the transaction probability as the weight of each node in the second transaction directed graph.

[0091] As one possible way to achieve this, such as Figure 6 As shown, based on the above implementation, the specific process of obtaining the transaction probability of each node in the second transaction directed graph in step S2010 includes steps S601 to S602:

[0092] S601, obtain the number of nodes in the second transaction directed graph, and initialize the first probability of each node having a transaction at time t based on the number of nodes.

[0093] It should be noted that if the number of nodes in the directed graph Graph2 of the second transaction is N, and each node has the same probability of being visited, then the first probability of each node having a transaction at time t is 1 / N.

[0094] S602, based on the first probability, iterate the second probability of each node having a transaction at time t+1. In response to the fulfillment of the iteration termination condition, take the second probability corresponding to the last iteration as the transaction probability of each node in the second transaction directed graph.

[0095] It should be noted that this disclosure does not limit the setting of the iteration termination condition, which can be set according to the actual situation. Optionally, the iteration termination condition is that the number of iterations is greater than or equal to a preset iteration number threshold; alternatively, the iteration termination condition is that the error value between the second probability and the first probability is less than or equal to a preset error threshold.

[0096] It should be noted that if the number of nodes in the second transaction directed graph Graph2 is N, at the initial time t=0, the probability that node A undergoes a transfer transaction (node A is transferred into) is P t=0 (A); at the next time t=1, the probability that node A performs a transfer transaction to node B (node A transfers out and node B is transferred into) is P t=1 (B|A), if the set of nodes that have transfer transactions with node A, where node A transfers out and the corresponding node is transferred into, is TransLink1(A), and the probability that node A transfers to each node in the set TransLink1(A) is the same, then where |TransLink1(A)| is the number of nodes in the set TransLink1(A); at the initial time t=0, node A has a transfer transaction, and the probability that node A performs a transfer transaction to node B at t=1 is P t=0 (A)*P t=1 (B|A); at t=1, the probability that node B undergoes a transfer transaction (node B is transferred into) is P t=1 (B)=∑ AinTransLink2(B) P t=0 (A)*P t=1 (B|A)=∑ AinTransLink2(B) P t=0 (A) / |TransLink1(A)| (Equation 1), where TransLink2(B) is the set of account nodes that have transfer transactions with node B, where account node B transfers in and the corresponding node transfers out; by generalizing Equation (1), that is, replacing the time t=0 with any time t=i, we can obtain:

[0097] P t=i+1 (B)=∑ AinTransLink2(B) P t=i (A) / |TransLink1(A)| (Equation 2)

[0098] For example, if the iteration termination condition is that the error value between the second probability and the first probability is less than a preset error threshold, calculate the second probability of each node performing a transaction at time t+1 according to Equation (2), sum the second probabilities of each node performing a transaction at time t+1, and when ∑ BinGraph2 P t=i+1 -∑ BinGraph2 P t=i <w, the iterative calculation is stopped. In this case, the second probability corresponding to the last iteration of each node is taken as the final transaction probability of each node in the second transaction directed graph, where B is all nodes in the second transaction directed graph, ∑ BinGraph2 P t=i+1 is the second probability, ∑ BinGraph2 P t=i is the first probability, and w is the error threshold.

[0099] In this embodiment of the disclosure, after obtaining the transaction probability of each node, the transaction probability is used as the weight wgt of each node in the second transaction directed graph. i .

[0100] S3011, divide the nodes in the second transaction directed graph according to the weights to obtain the node set of the second transaction directed graph.

[0101] As one possible way to achieve this, such as Figure 7 As shown, based on the above implementation, the specific process of dividing the nodes in the second transaction directed graph according to weights in step S2011 to obtain the node set of the second transaction directed graph includes steps S701 to S703:

[0102] S701, obtain the key nodes and important nodes of the key nodes in the directed graph of the second transaction.

[0103] In this embodiment of the disclosure, a preset node quantity threshold minCount and weight threshold minWgt are obtained for the node set. The transaction path length between node i and node j is obtained. Based on the transaction path length, the set of neighboring nodes of node i in the second transaction directed graph is determined. In response to the number of nodes in the set of neighboring nodes of node i being greater than or equal to the node quantity threshold and the weight sum of the set of neighboring nodes of node i being greater than or equal to the weight threshold, node i is determined to be a key node. In response to node k being located in the set of neighboring nodes of node i and node i being a key node, node k is determined to be an important node of node i.

[0104] For example, for nodes A and B, if the number of nodes in the neighbor set of node A is greater than or equal to the node count threshold minCount, and the weights and values ​​of the neighbor set of node A are... If node A is greater than or equal to the weight threshold minWgt, then node A is a critical node; if node B is in the set of neighboring nodes of node A, and node A is a critical node, then node B is an important node of node A.

[0105] In this embodiment of the disclosure, the maximum number of edges from node i to node j is determined in the second directed transaction graph. Based on the maximum number of edges, the transaction path length between node i and node j is determined, and a preset transaction path length threshold len is obtained. In response to the transaction path length between node i and node j being less than or equal to the transaction path length threshold, node j is designated as a neighbor node of node i in the second directed transaction graph, wherein the set of neighbor nodes includes all neighbor nodes of node i.

[0106] For example, such as Figure 5As shown, nodes A1 and A2 have a transaction, and nodes A2 and A4 have a transaction. Therefore, the transaction path length between nodes A1 and A2 is 2.

[0107] S702, divide each key node and the important nodes of each key node into the candidate node set, and determine whether there are duplicate nodes in the candidate node set.

[0108] S703, in response to the existence of identical nodes in the candidate node set, merge the candidate node sets until there are no identical nodes in the merged candidate node set, and obtain the node set in the second transaction directed graph.

[0109] For example, if node A is a critical node, nodes B and C are important nodes of node A, and account nodes B and C are also critical nodes, node D is an important node of node B, and node E is an important node of node C, then nodes A, B, C, D, and E are merged and assigned to the same node set.

[0110] S3012, obtain the target transaction label prediction value of each node in the node set, and determine the second identification result of the second transaction data based on the target transaction label prediction value.

[0111] As one possible way to achieve this, such as Figure 8 As shown, based on the above implementation, the specific process of obtaining the target transaction tag prediction value of each node in the node set in step S2012, and determining the second identification result of the second transaction data based on the target transaction tag prediction value, includes steps S801 to S805:

[0112] S801, retrieve the target transaction data corresponding to each node in the node set.

[0113] In this embodiment of the disclosure, after obtaining the set of nodes in the second directed graph of transactions, the target transaction data corresponding to each node in the set of nodes can be obtained.

[0114] S802: Obtain the target feature vector of the target transaction data, input the target feature vector into the pre-trained supervised transaction recognition model, and obtain the target transaction label prediction value.

[0115] In this disclosed embodiment, after obtaining the target transaction data, feature extraction can be performed on the target transaction data to obtain the target feature vector of the target transaction data. The target feature vector is then input into a pre-trained supervised transaction recognition model, which outputs the predicted value y of the target transaction label. n .

[0116] S803, accumulate the predicted values ​​of the target transaction labels to obtain the predicted sum of the target transaction labels for the node set.

[0117] For example, given the node set {A} in the directed graph of the second transaction, if the node set {A} includes node A... 1 Node A 2 and node A 3 Node A 1 Node A 2 and node A 3 If the predicted values ​​of the target transaction labels are y1, y2, and y3, then the predicted sum of the target transaction labels for the node set {A} is y1, y2, and y3. sum = y1 + y2 + y3.

[0118] S804, in response to the target transaction label prediction and value being less than a preset second threshold, determines that all transaction data in the node set are abnormal.

[0119] It should be noted that this disclosure does not impose any restrictions on the setting of the second threshold bad_score, which can be set according to the actual situation.

[0120] For example, if the set of nodes {A} includes node A 1 A 2 and A 3 The predicted target transaction label and value y of the node set {A} sum If the target transaction label prediction and value y sum If the score is less than the preset second threshold "bad_score", then all transaction data in the node set is determined to be abnormal, i.e., node A. 1 Node A 2 and node A 3 All transaction data is abnormal.

[0121] S805, in response to the target transaction label prediction and value being greater than or equal to a preset second threshold, determines that all transaction data in the node set are normal.

[0122] For example, if the set of nodes {A} includes node A 1 A 2 and A 3 The predicted target transaction label and value y of the node set {A} sum If the target transaction label prediction and value y sum If the score is greater than or equal to the preset second threshold bad_score, it is determined that all transaction data in the node set is normal, i.e., node A... 1 Node A 2 and node A 3 All transaction data are normal.

[0123] S3013, Based on the second identification result, filter out the third transaction data that is abnormal from the second transaction data.

[0124] In this embodiment of the disclosure, after obtaining the second identification result, third transaction data that is abnormal can be filtered from the second transaction data.

[0125] In summary, according to the transaction identification method provided in this disclosure, a first directed transaction graph is constructed based on transaction accounts and the transaction transfer relationships between every two transaction accounts. The nodes of the first directed transaction graph represent transaction accounts, and the edges represent the transaction transfer relationships between every two transaction accounts. A supervised transaction identification model is obtained by supervising the initial transaction identification model using a training sample set. The supervised transaction identification model is then used to identify the first transaction data, determining the normal transaction data within it. Based on the normal transaction data, the first directed transaction graph is pruned to obtain a second directed transaction graph. Finally, the nodes in the second directed transaction graph are further processed... The method involves dividing the data into rows to obtain a set of nodes in a directed graph of the second transaction. Based on the predicted value of the target transaction label for each node in the set, a second identification result for the second transaction data is determined. Based on the second identification result, third transaction data with anomalies is filtered out from the second transaction data. Thus, this method can identify a large amount of transaction data, improve the efficiency of obtaining abnormal transaction data, reduce the time spent on manual monitoring and identification of abnormal transaction data, avoid the problems of oversight and misjudgment in the transaction data identification process caused by subjective factors, and ensure the objectivity of the transaction identification results by not relying on pre-set rules, thereby improving the accuracy and reliability of obtaining abnormal transaction identification data.

[0126] Figure 9 This is a block diagram of a transaction identification device according to a first embodiment of the present disclosure.

[0127] like Figure 9 As shown, the transaction identification device 900 of this embodiment includes: a first acquisition module 901, a second acquisition module 902, a determination module 903, a cropping module 904, and an identification module 905.

[0128] The first acquisition module 901 is configured to acquire the first transaction data to be identified, and to acquire the transaction accounts and the transaction transfer relationships between every two transaction accounts from the first transaction data;

[0129] The second acquisition module 902 is configured to construct a first directed transaction graph based on the transaction account and the transaction transfer relationship between every two transaction accounts;

[0130] The determination module 903 is configured to identify the first transaction data, obtain a first identification result of the first transaction data, and determine normal transaction data in the first transaction data based on the first identification result;

[0131] The trimming module 904 is configured to trim the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph;

[0132] The identification module 905 is configured to determine the second transaction data based on the nodes in the second transaction directed graph, and to identify the second transaction data in order to filter out the third transaction data that is abnormal from the second transaction data.

[0133] In one embodiment of this disclosure, the trimming module 904 is further configured to: obtain a first transaction account and a second transaction account that have a transaction transfer relationship in the normal transaction data; determine a first node associated with the first transaction account and a second node associated with the second transaction account in the first directed transaction graph; and delete the edge between the first node and the second node to obtain the second directed transaction graph.

[0134] In one embodiment of this disclosure, the identification module 905 is further configured to: obtain the transaction probability of each node in the second directed transaction graph, and use the transaction probability as the weight of each node in the second directed transaction graph; divide the nodes in the second directed transaction graph according to the weight to obtain a node set of the second directed transaction graph; obtain the target transaction label prediction value of each node in the node set, and determine the second identification result of the second transaction data according to the target transaction label prediction value; and filter out the third transaction data with anomalies from the second transaction data according to the second identification result.

[0135] In one embodiment of this disclosure, the identification module 905 is further configured to: obtain the number of nodes in the second transaction directed graph; initialize a first probability of each node conducting a transaction at time t based on the number of nodes; iterate the second probability of each node conducting a transaction at time t+1 based on the first probability; and, in response to satisfying the iteration termination condition, use the second probability corresponding to the last iteration as the transaction probability of each node in the second transaction directed graph.

[0136] In one embodiment of this disclosure, the identification module 905 is further configured to: acquire key nodes and important nodes of the key nodes in the second transaction directed graph; divide each key node and each important node of the key node into a candidate node set, and determine whether there are identical nodes in the candidate node set; in response to the existence of identical nodes in the candidate node set, merge the candidate node set until there are no identical nodes in the merged candidate node set, thereby obtaining a node set in the second transaction directed graph.

[0137] In one embodiment of this disclosure, the identification module 905 is further configured to: obtain a preset node quantity threshold and weight threshold for the node set; obtain the transaction path length between node i and node j, and determine the set of neighboring nodes of node i in the second directed transaction graph based on the transaction path length; determine node i as a key node in response to the number of nodes in the set of neighboring nodes of node i being greater than or equal to the node quantity threshold and the sum of the weights of the set of neighboring nodes of node i being greater than or equal to the weight threshold; and determine node k as an important node of node i in response to node k being located in the set of neighboring nodes of node i and node i being a key node.

[0138] In one embodiment of this disclosure, the identification module 905 is further configured to: determine the maximum number of edges from node i to node j in the second directed transaction graph; determine the transaction path length between node i and node j based on the maximum number of edges and obtain a preset transaction path length threshold; and, in response to the transaction path length between node i and node j being less than or equal to the transaction path length threshold, designate node j as a neighbor node of node i in the second directed transaction graph, wherein the set of neighbor nodes includes all neighbor nodes of node i.

[0139] In one embodiment of this disclosure, the identification module 905 is further configured to: acquire target transaction data corresponding to each node in the node set; acquire the target feature vector of the target transaction data, input the target feature vector into a pre-trained supervised transaction identification model to obtain the target transaction label prediction value; accumulate the target transaction label prediction values ​​to obtain the target transaction label prediction sum value of the node set; in response to the target transaction label prediction sum value being less than a preset second threshold, determine that all transaction data in the node set are abnormal; or, in response to the target transaction label prediction sum value being greater than or equal to the preset second threshold, determine that all transaction data in the node set are normal.

[0140] In one embodiment of this disclosure, the determining module 903 is further configured to: acquire a feature vector of the first transaction data, input the feature vector into a pre-trained supervised transaction recognition model to obtain a first transaction label prediction value of the first transaction data; determine that the first transaction data is normal transaction data in response to the first transaction label prediction value being greater than a preset first threshold; or determine that the first transaction data is abnormal transaction data in response to the first transaction label prediction value being less than or equal to a preset first threshold.

[0141] In one embodiment of this disclosure, the training process of the supervised transaction recognition model includes: acquiring a training sample set, wherein each training sample in the training sample set includes sample transaction data and sample transaction label real values; acquiring sample feature vectors of the sample transaction data, inputting the sample feature vectors into the initial transaction recognition model, and acquiring transaction label prediction values; and performing supervised training on the initial transaction recognition model based on the sample transaction label real values ​​and the transaction label prediction values ​​to obtain the supervised transaction recognition model.

[0142] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0143] In summary, the transaction identification device provided in this disclosure acquires first transaction data to be identified, obtains transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data, constructs a first directed transaction graph based on the transaction accounts and transaction transfer relationships between every two transaction accounts, identifies the first transaction data, obtains a first identification result of the first transaction data, determines normal transaction data in the first transaction data based on the first identification result, trims the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph, determines second transaction data based on the nodes in the second directed transaction graph, and identifies the second transaction data to filter out abnormal third transaction data from the second transaction data. Therefore, this disclosure can identify massive amounts of transaction data, improves the efficiency of identifying abnormal transaction data, ensures the objectivity of obtaining abnormal transaction identification data, saves costs and time in the transaction identification process, and improves the accuracy and reliability of obtaining abnormal transaction identification data.

[0144] Figure 10 This is a block diagram illustrating an electronic device according to an exemplary embodiment.

[0145] like Figure 10 As shown, the aforementioned electronic device 1000 includes:

[0146] The memory 010 and processor 1020 are connected by a bus 1030, which connects different components (including the memory 1010 and the processor 1020). The memory 1010 stores a computer program, which implements the transaction identification method described in this embodiment when the processor 1020 executes the program.

[0147] Bus 1030 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0148] Electronic device 1000 typically includes a variety of electronic device readable media. These media can be any available media that can be accessed by electronic device 1000, including volatile and non-volatile media, removable and non-removable media.

[0149] The memory 1010 may also include computer system readable media in the form of volatile memory, such as random access memory (RAM) 1040 and / or cache memory 1050. The electronic device 1000 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 1060 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 10 Not shown; usually referred to as a "hard drive"). Although Figure 10 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 1030 via one or more data media interfaces. Memory 1010 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this disclosure.

[0150] A program / utility 1080 having a set (at least one) of program modules 1070 may be stored, for example, in memory 1010. Such program modules 1070 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 1070 typically perform the functions and / or methods described in the embodiments of this disclosure.

[0151] Electronic device 1000 can also communicate with one or more external devices 1090 (e.g., keyboard, pointing device, display 1091, etc.), and with one or more devices that enable a user to interact with the electronic device 1000, and / or with any device that enables the electronic device 1000 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 1092. Furthermore, electronic device 1000 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1093. Figure 10 As shown, network adapter 1093 communicates with other modules of electronic device 1000 via bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0152] The processor 1020 performs various functional applications and data processing by running programs stored in the memory 1010.

[0153] It should be noted that the implementation process and technical principles of the electronic device in this embodiment are explained in the foregoing description of the transaction identification method of this disclosure embodiment, and will not be repeated here.

[0154] In summary, the electronic device provided in this disclosure can execute the transaction identification method described above. It acquires first transaction data to be identified, obtains transaction accounts and transaction transfer relationships between every two transaction accounts from the first transaction data, constructs a first directed transaction graph based on the transaction accounts and transaction transfer relationships between every two transaction accounts, identifies the first transaction data, obtains a first identification result of the first transaction data, determines normal transaction data in the first transaction data based on the first identification result, trims the first directed transaction graph based on the normal transaction data to obtain a second directed transaction graph, determines second transaction data based on the nodes in the second directed transaction graph, and identifies the second transaction data to filter out abnormal third transaction data from the second transaction data. Therefore, this disclosure can identify massive amounts of transaction data, improves the efficiency of identifying abnormal transaction data, ensures the objectivity of obtaining abnormal transaction identification data, saves costs and time in the transaction identification process, and improves the accuracy and reliability of obtaining abnormal transaction identification data.

[0155] To implement the above embodiments, this disclosure also proposes a computer-readable storage medium.

[0156] When the instructions in the computer-readable storage medium are executed by the processor of an electronic device, the electronic device is able to perform the transaction identification method as described above. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc.

[0157] To implement the above embodiments, this disclosure also provides a computer program product, including a computer program, characterized in that, when the computer program is executed by a processor, it implements the transaction identification method as described above.

[0158] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0159] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A transaction identification method, characterized in that, The method includes: Obtain the first transaction data to be identified, and obtain the transaction accounts and the transaction transfer relationships between every two transaction accounts from the first transaction data; Based on the transaction accounts and the transaction transfer relationships between every two transaction accounts, a first directed transaction graph is constructed; The first transaction data is identified to obtain a first identification result of the first transaction data, and normal transaction data in the first transaction data is determined based on the first identification result. The first directed graph of transactions is cropped based on the normal transaction data to obtain the second directed graph of transactions; Based on the nodes in the second directed graph of the transaction, the second transaction data is determined and identified in order to filter out the third transaction data that is abnormal from the second transaction data; The step of trimming the first directed graph based on the normal transaction data to obtain the second directed graph includes: Obtain the first and second transaction accounts that have a transaction transfer relationship from the normal transaction data; In the first directed graph of transactions, determine the first node associated with the first transaction account and the second node associated with the second transaction account; Delete the edge between the first node and the second node to obtain the second transaction directed graph; The step of identifying the second transaction data to filter out third transaction data that is abnormal includes: Obtain the transaction probability of each node in the second directed transaction graph, and use the transaction probability as the weight of each node in the second directed transaction graph; The nodes in the second directed graph of transactions are divided according to the weights to obtain the node set of the second directed graph of transactions; Obtain the target transaction tag prediction value for each node in the node set, and determine the second identification result of the second transaction data based on the target transaction tag prediction value; Based on the second identification result, filter out the third transaction data that contains anomalies from the second transaction data.

2. The identification method according to claim 1, characterized in that, The step of obtaining the transaction probability of each node in the second transaction directed graph includes: Obtain the number of nodes in the second transaction directed graph, and initialize each node according to the number of nodes. t The first probability of a transaction occurring at any given moment; Based on the first probability, for each node in t The second probability of a transaction occurring at time +1 is iterated. In response to the fulfillment of the iteration termination condition, the second probability corresponding to the last iteration is taken as the transaction probability of each node in the second transaction directed graph.

3. The identification method according to claim 1, characterized in that, The step of partitioning the nodes in the second directed graph according to the weights to obtain the node set of the second directed graph includes: Obtain the key nodes and important nodes of the second transaction directed graph; Each key node and its important nodes are assigned to a candidate node set, and it is determined whether there are any identical nodes in the candidate node set. In response to the presence of identical nodes in the candidate node set, the candidate node set is merged until no identical nodes exist in the merged candidate node set, thus obtaining the node set in the second transaction directed graph.

4. The identification method according to claim 3, characterized in that, The process of obtaining the key nodes and important nodes of the second transaction directed graph includes: Obtain the preset threshold for the number of nodes and the threshold for the weight of the node set; Get Nodes i With nodes j The length of the transaction path between them is determined, and based on the length of the transaction path, the nodes in the second transaction directed graph are determined. i The set of neighboring nodes; Response to the node i The number of nodes in the neighbor set is greater than or equal to the node number threshold and the nodes i If the sum of the weights of the neighboring nodes is greater than or equal to the weight threshold, then the node is determined. i As a key node; Response to node k Located at node i In the set of neighboring nodes and nodes i As a key node, determine the node. k For the node i A crucial juncture.

5. The identification method according to claim 4, characterized in that, The acquisition node i With nodes j The length of the transaction path between them is determined, and based on the length of the transaction path, the nodes in the second transaction directed graph are determined. i The set of neighboring nodes includes: Determine the slave node in the second transaction directed graph. i To the node j The maximum number of edges; Determine the node based on the maximum number of edges. i With nodes j Transaction path length between Obtain the preset transaction path length threshold; Response to the node i With nodes j If the transaction path length between nodes is less than or equal to the transaction path length threshold, the nodes will be... j As a node in the second transaction directed graph i The set of neighboring nodes includes nodes. i All neighboring nodes.

6. The identification method according to claim 1, characterized in that, The step of obtaining the target transaction tag prediction value for each node in the node set, and determining the second identification result of the second transaction data based on the target transaction tag prediction value, includes: Obtain the target transaction data corresponding to each node in the node set; Obtain the target feature vector of the target transaction data, and input the target feature vector into a pre-trained supervised transaction recognition model to obtain the predicted value of the target transaction label; The predicted values ​​of the target transaction tags are summed to obtain the predicted sum of the target transaction tags for the node set. In response to the predicted value of the target transaction label being less than a preset second threshold, it is determined that all transaction data in the node set are abnormal; or... In response to the predicted sum of the target transaction tags being greater than or equal to a preset second threshold, it is determined that all transaction data in the node set are normal.

7. The identification method according to claim 1, characterized in that, The step of identifying the first transaction data and obtaining a first identification result of the first transaction data includes: Obtain the feature vector of the first transaction data, and input the feature vector into a pre-trained supervised transaction recognition model to obtain the first transaction label prediction value of the first transaction data; In response to the first transaction tag prediction value being greater than a preset first threshold, the first transaction data is determined to be normal transaction data; or, In response to the first transaction tag prediction value being less than or equal to a preset first threshold, the first transaction data is determined to be abnormal transaction data.

8. The identification method according to claim 7, characterized in that, The training process of the supervised transaction recognition model includes: Obtain a training sample set, wherein each training sample in the training sample set includes sample transaction data and the true value of sample transaction label; Obtain the sample feature vector of the sample transaction data, input the sample feature vector into the initial transaction recognition model, and obtain the transaction label prediction value; Based on the actual values ​​of the sample transaction labels and the predicted values ​​of the transaction labels, the initial transaction recognition model is trained in a supervised manner to obtain the supervised transaction recognition model.

9. A transaction identification device, characterized in that, The device includes: The first acquisition module is configured to acquire the first transaction data to be identified, and to acquire the transaction accounts and the transaction transfer relationships between every two transaction accounts from the first transaction data; The second acquisition module is configured to construct a first directed transaction graph based on the transaction account and the transaction transfer relationship between every two transaction accounts; The determination module is configured to identify the first transaction data, obtain a first identification result of the first transaction data, and determine normal transaction data in the first transaction data based on the first identification result. The trimming module is configured to trim the first directed graph of transactions based on the normal transaction data to obtain a second directed graph of transactions; The identification module is configured to determine the second transaction data based on the nodes in the second transaction directed graph, and to identify the second transaction data in order to filter out the third transaction data that is abnormal from the second transaction data; The trimming module is configured to: obtain a first transaction account and a second transaction account that have a transaction transfer relationship in the normal transaction data; determine a first node associated with the first transaction account and a second node associated with the second transaction account in the first directed transaction graph; and delete the edge between the first node and the second node to obtain the second directed transaction graph. The identification module is configured to: obtain the transaction probability of each node in the second directed transaction graph, and use the transaction probability as the weight of each node in the second directed transaction graph; divide the nodes in the second directed transaction graph according to the weight to obtain the node set of the second directed transaction graph; obtain the target transaction label prediction value of each node in the node set, and determine the second identification result of the second transaction data according to the target transaction label prediction value; and filter out the third transaction data with anomalies from the second transaction data according to the second identification result.

10. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the transaction identification method as described in any one of claims 1-8.

11. A computer-readable storage medium, wherein instructions in the computer-readable storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the transaction identification method as described in any one of claims 1-8.

12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the transaction identification method as described in any one of claims 1-8.

Citation Information

Patent Citations

  • Method and device for identifying associated specific account and storage medium

    CN111831923A

  • Network training method, abnormal transaction behavior identification method and device, and medium

    CN112435122A